Is legalese not just the result of trying to use English as a programming language? Any time I try to write English (or other natlang) precisely and unambiguously and robust against adversarial interpretations, it comes out looking like legalese.
The thinking being that the less nuanced the vocabulary, the less ways it could be interpreted and thus one may not have to write so many laywerly guard phrases to artificially constrain "normal" vocabulary. It may very well run the risk of having to use a bazillion more cross-references as one builds up a "library" of word-subroutines, but still could be a net win
But I guess I can navel-gaze all I like because for this specific domain, any change might as well be all the changes since there's no prayer
The plainlanguage.gov site is an excellent all-around writing resource. I direct junior developers here when they are trying too hard to sound fancy when communicating technical concepts in documentation and design documents.
Here are some great examples:
https://www.plainlanguage.gov/examples/before-and-after/ambi...
https://www.plainlanguage.gov/examples/before-and-after/mont...
https://www.plainlanguage.gov/examples/before-and-after/use-...
well, no shit! that's amazing
Thanks so much for bringing that to my attention, I'll try to see how I can incorporate those into my own process
As a particular example, is there any reason to keep the vague "second month" in the second example [2], rather than "subsequent month" or "next month"?
[2] https://www.plainlanguage.gov/examples/before-and-after/mont...
My friends from non English countries get very confused that somehow “next Wednesday” when it’s Monday might not mean two days from now but 8 days from now. And how two days in that instance would be referred as “this Wednesday” or “this coming Wednesday”
Which is different way of talking. If you were sitting by the road counting cars and you are at car “n”, Saying “the next car” would refer to car n+1. If your counting wednesdays you experienced “next Wednesday” technically refers to n+2
I stopped saying “this <day>” or “next <day>” and now just say “Wednesday the 25th” for instance.
'Next Wednesday' is always the Wednesday of the calendar week following the current calendar week; doesn't matter what day of the week it currently is. 'This Wednesday' is always the Wednesday of current week on the calendar— even if that day is in the past.
Is it quirky that this expression doesn't instead mean 'the next Wednesday that will occur'? Yes, definitely. But I don't see how it's difficult to describe what it actually does mean.
> I stopped saying “this <day>” or “next <day>” and now just say “Wednesday the 25th” for instance.
I love this. Indexicals in general can be tricky, and I love expressions that rely less or very little at all on context. Sometimes when a friend is telling a complicated story I'll ask them to repeat something tbey just said but with no use of pronouns, for instance, and it always makes interpretation much easier.
As much as I think the actual idioms are perfectly describable, they are somewhat prominently misused. One of my pet peeves is how YouTube's search filters uses its time restriction phrases incorrectly: it says 'today' to mean 'within the past 24 hours', 'this week' to mean 'within the past week', 'this year' to mean 'within the past year', etc. It's Tuesday, and when I search for videos with an upload date from 'this week', I get results including videos uploaded 4 days ago, but this week is not yet 4 days old under any standard convention (e.g., starting the week on Mondays rather than Sundays)... -_-
'<term>' is always <my definition>; doesn't matter <other factor>.
Is never going to be true in spoken language. Otherwise we wouldn't be having conversations about confusion and ambiguity in the first place.
By my intuition, I'd say the uses of 'this week' and friends I described as improper qualify handily, not being all that widespread yet. A couple more marginal examples that stand out to me because erroneous uses are much more common: 'let alone' (which is binary and often used with subject and object reversed), the distinction between envy and jealousy.
You can take up a radical descriptivist position, e.g., that anything spoken by adult native speakers of normal faculties is never erroneous, or re-scope my assertions by saying that I'm gatekeeping speakers who see such usage as correct from my perceived language community or tradition, whatever, but imo the first is trivial and the second boring.
Sometimes specialists misappropriate methodological constraints from their discipline as general ontological or social principles, often discounting an inherent normativity in the way people actually relate to the things those specialists study. I think that's essentially at the heart of the most inflated and controversial uses of concepts like cultural relativism and linguistic descriptivism, and probably applies to nonspecific objections like the one you make above (as opposed to pointing at some specific dialectal variation in the use of such phrases or something like that).
Nope. This is highly contingent on "which day of the week starts a week" conventions.
If it is Sunday, then "next Wednesday" does not clearly mean the day 10 days from now ... and which one you lean toward will depend (in part) on "week starts on Monday" or "week starts on Sunday".
Things are also a bit complicated where I live by the fact that 'this weekend' and 'next weekend' follow the same pattern as I described before, but in a way consistent with calendar weeks beginning on Monday rather than Sunday— even though calendars here conventionally start the week on Sunday and usage of 'this <day of week>' and 'next <day of week>' align with that.
Anyhow, the variation you are getting at is already captured in the description I gave: as the calendar (week boundaries) varies, so does the description's meaning. The description is already indexed to a particular calendar (as is the expression, unfortunately implicitly). :p
My claim was that the meaning is easy enough to describe, not that the phrase is unambiguous. That a phrase can be used ambiguously doesn't mean that descriptions/definitions/characterizations of its general meaning actually have to be ambiguous or vague themselves.
Still yeah, this is a real problem for conversations between people who aren't looking at the same calendar/don't understand a shared convention for week boundaries.
My sense is that weekend nomenclature is even more confusing.
If it is Monday, and I say "next weekend", I'd wager there's a greater proportion of English speakers (at least) who would understand that to mean the two day period that starts in roughly another 4 days. That is: "next weekend" referred to at any time before (possibly) Friday means "the next one to occur", not "the one that is a part of the next calendar week". By contrast, on Friday "next weekend" pretty clearly means the two days that will occur in about 7 days, rather than "this weekend" meaning the two days that start in less than 24 hours.
> If it is Monday, and I say "next weekend", I'd wager there's a greater proportion of English speakers (at least) who would understand that to mean the two day period that starts in roughly another 4 days.
To my ear, this is wrong, but I think you're right that it's more common. I think 'this weekend' and 'next weekend' are fuzzier than 'this Tuesday' and 'next Tuesday' because of that misalignment I pointed to before, where 'the weekend' is thought of as a thing that comes at the end of a week, but how we write it most calendars where I live is as a thing that bookends the week on each side.
But I couldn't take up the usage you describe even if I moved to a region where it was predominant, I think. How could I abide a situation where 'this Saturday' occurs during 'next weekend'? That's simply madness. :D
(It's amazing how much we humans manage to communicate with something as messy as natural language-- perhaps especially amazing to programmer-brained people like me, who take some comfort and ease from the simplicity and neatness of the artificial formalisms we work with every day.)
for weekends, most of the time most people are talking in ways that make "this weekend" "the next weekend to occur". so here, "this" and "next" are effectively synonymous.
for days of the week, however, it is more often important to differentiate "this Thursday" (occuring in a couple of days from now, during this week) and "next thursday" (occuring in more than 7 days from now, as part of next week).
If my hunch is correct, it would make conventions highly dependent on social scheduling frequency in a culture.
Is it quirky that this expression doesn't instead mean 'the next Wednesday that will occur'? Yes, definitely. But I don't see how it's difficult to describe what it actually does mean.
That's the way it's supposed to be, at least as spoken in the part of Scotland from where my family hails.
But the Americans I know seem to get it wrong about half the time.
What I haven't figured out is if it's a regional (dialectical) thing, or just certain people being "dense", or simply never having been taught the rule.
> Sid: Well I'm going down to visit my sister in Virginia next Wednesday, for a week, so I can't park it.
> Jerry: This Wednesday?
> Sid: No, next Wednesday, week after this Wednesday.
> Jerry: But the Wednesday two days from now is the next Wednesday.
> Sid: If I meant this Wednesday, I would have said this Wednesday. It's the week after this Wednesday.
Well, usually. I've still had conversations like this:
Me: "It'll be next week on Wednesday."
Them: "Whoa, slow down there. Do you mean this coming Wednesday, or next week on Wednesday?"
Me: "It'll be next week on Wednesday."
Them: "Ohhhh! Thank you so much, it wasn't clear before."
If they name a rule after me, I want it to be, "If someone can reasonably interpret a phrase to mean two contradictory things, you have to stop using that phrase if you want clear communication, even if it means being more verbose."
I can easily say "Hey, let's meet next Wednesday!" and you may reply with "Oh, great, I'll put it in the calendar, the 8th it is!" and you reply with "No, I meant the 15th" -- and this reduces the ambiguity. This kind of thing would be happening in "real time", so being perfectly clear isn't as important.
However, if you write "next Wednesday" in a contract, and then complete it on the 15th instead of the 8th, and it's taken to Court ... it's a lot harder to figure out what the the original conversation was like, that led to the writing of the contract!
Hence, it makes sense to use as unambiguous language as possible when writing a contract.
Maybe because I’m on mobile?
I genuinely don't mean this in a dickish way -- isn't this, like, tautologically untrue?
By definition, more nuanced, more descriptive language describes a narrower, more precise view of reality than broader language otherwise would.
When would plainer language allow less room for interpretation?
I do generally think writing laws and other documents in plainer language would be beneficial for society, but not for this reason. Sometimes you do have to describe a really, really precise concept. "Kill" is different than "murder" is different than "manslaughter" in ways that are meaningful and important to preserve.
Although even as I write that, I guess you could say "kill", "kill a person with intention", "kill a person without intention". That's kind of what you mean by word subroutines?
At a certain point this just seems like a similarly-complex vocabulary, just with more words, though.
I am 100% open to the fact that it may not be possible to do this, since nat-lang is its own special little thing, and trying to apply fixes to it may be nonsensical themselves
The word subroutines would be cross-references to potentially more complex concepts akin to "one cannot end life (§3.14.159) unless working (§8.6.753) in a job (§127.0.1) that allows State violence" where the boundaries of what this legislation cares about 'ending life,' the boundaries around 'working,' the boundaries of a 'job' would then be composed into 'citizen cannot kill other citizen'.
I always got the impression that the nuance between murder and manslaughter wasn't in their degree of unlawfulness but rather in their sentencing, but I am deeply thankful that I haven't needed to know
As an aside the difference between murder and manslaughter is in the intent of the perpetrator. Murder is typically when you intended for the outcome to be death (and is additionally divided into whether or not it was premeditated/planned).
Manslaughter is reserved for when there was not intent to kill, but your actions caused a death.
Err, having written that out I now guess there is also some social component to it: you may still be received at a party if convicted of manslaughter but maybe not murder so we need different words to describe the act for purposes outside of the legal system
We even draw the distinction between degrees of murder since sitting down and planning a murder in cold blood (murder in the first degree) is far different than a road rage incident with a gun (murder 2) which is different still than a shove in a bar where someone falls down and hits their head and dies (manslaughter). Hell, some places even distinguish between voluntary and involuntary manslaughter.
The point is that all these words have meaning, and we deeply care about the nuance.
To circle back to the "if law were programming" idea, I think of all the nuances you cited as belonging in any "then" clauses, not the "if" clauses of legalese
if made_dead:
# determine punishment
if premeditated:
...
elif road_rage and crime_scene.contains({"gun"}):
# a kind of very, very slightly premeditation
...
else:
# and here one can get into [in]voluntary made dead
All this was inscrutable before LLMs, but LLMs bring their own challenges: to summarize something in plain text, is it using a deep graph of definitions that are sourced and verifiable, or hallucinating their existence? IMO architectures as in https://arxiv.org/html/2410.04949v1 and https://arxiv.org/html/2409.13252v1 are useful; one uses LLMs to create local knowledge graphs and integrate them, then translates natural language queries into (successive) graph queries or graph-based RAG approaches. Things are still evolving in real time here, and IMO we've only scratched the surface of what's possible.
You want to administer nuclear weapons, the U.S. military and toxic-waste rules based on a high-school freshman’s knowledge of the world?
I think that's why these terms exist, because they become shorthand for longwinded definitions that may need to be very precise.
Real life can be complicated and moral/legal questions can be hard to determine.
Seems like there's a lot of pitfalls there, but that comes with the territory of writing laws in general.
Seems like a concept worth exploring.
I am pretty sure there are several more but that's the one I can recall. It has been discussed here periodically https://news.ycombinator.com/from?site=github.com/catalalang (and https://news.ycombinator.com/from?site=catala-lang.org ) with the 2023 thread being the beefiest https://news.ycombinator.com/item?id=37546874
If we were going to get 80 year old Senators to learn a new language, I'd lobby for Lojban which is at least plausibly designed for representing unambiguous semantics
I've mentioned before that any laws that are written in formal languages seems great from the metric of keeping "interpretation" cases out of the judicial system but would put society back into the "priests read Latin and tell the plebs what God really meant, trustmebro" and that's for sure no bueno
For instance, here [1] is a random paragraph I found in a contract that I think is pretty good example of "legalese", and here [2] is my attempt to rewrite it for readability. All the essential complexity remains, but I think (hope!) much of the accidental complexity has been removed. :)
[1] 3.3.4 Date of Issuance. Each person in whose name any book entry position or certificate for shares of Common Stock is issued shall for all purposes be deemed to have become the holder of record of such shares on the date on which the Warrant, or book entry position representing such Warrant, was surrendered and payment of the Warrant Price was made, irrespective of the date of delivery of such certificate, except that, if the date of such surrender and payment is a date when the stock transfer books of the Company or book entry system of the Warrant Agent are closed, such person shall be deemed to have become the holder of such shares at the close of business on the next succeeding date on which the stock transfer books or book entry system are open.
[2] 3.3.4 Date of Issuance. To determine the record date for ownership of Common Stock shares (whether issued as a book entry or certificate), ask: Were the Company's stock transfer books and the Warrant Agent's book entry system open when the Warrant was surrendered and the Warrant Price was paid? If yes, the record date is that same date of surrender and payment. If no, the record date is the close of business on the next day when the books and systems are open.
I suspect you will end up with something similar to the original.
In your rewrite, if both the stock transfer books and the book entry system of the warrant agent are closed, the person becomes a holder on the next day when both are open.
If you search for the language of the original, you'll find a bunch of examples of the exact same language. I'm with the others that this is well-litigated language that no one wants to change. https://www.bing.com/search?pglt=161&q=Each+person+in+whose+....
[2] 3.3.4 Date of Issuance. To determine the record date for ownership of Common Stock shares (whether issued as a book entry or certificate), ask: Were the Company's stock transfer books or the Warrant Agent's book entry system open when the Warrant was surrendered and the Warrant Price was paid? If yes, the record date is that same date of surrender and payment. If no, the record date is the close of business on the next day when either the books or systems are open.
Been there. Learned eventually. Sometimes still forget. :)
Who told you forum shopping is illegal to talk about?
All this stuff is hard to navigate if you're not used to it, or haven't been involved before.
Edit: Actually we do. Skilled interlocutors like that doing their thing are how we got leaded gasoline.
This sounds like something a lawyer would say to a client who wants to think that. Law and coding have remarkable parallels.
True. But this isn’t because someone is more logical. Honestly, that was a great line by a lawyer who probably wanted to focus on the case and not bill hours for a philosophy of law discussion.
Not really. In particular, they’re both professions filled with people who have egos the size of planets. I can just as easily see a surgeon telling a lawyer that the law is logical, being designed by man, in a way the human body is not just to get them to shut the hell up with broad questions about human anatomy during a surgical consult.
(The actual parallel is they both deal with constructed languages. High-level languages are full of hacks and quirks and high-octane stupid it, just like the law.)
This isn't necessarily a bad thing. A more precisely worded contract, for example, is arguably more likely to have unambiguous loopholes that people can abuse without you being able to easily fight back. The well-known reductio ad absurdum of this phenomenon is Etherium smart contracts.
You see this in laws, too. The US's Federal Rules of Civil Procedure and associated case law, for example, contain all sorts of explicit refusals to say things more precisely. The stated rationale, here, is that it's impossible for the people drafting these rules to anticipate every possible situation and contingency, and instead they must trust that reasonable attorneys and judges are able sort things out in the course of litigation.
The article then went on to say this isn't necessarily a bad thing: two lawyers put their arguments before the judge, and the judge can then decide what's best for that particular situation.
The author of the article was at least somewhat libertarian, because he suggested that this is necessary when there's a single body of Law to deal with; the other way to deal with these problems is to go with arbitrators instead -- but in that case, the "Law" would be decided between the two parties and the arbitrator, which has its own twists and turns!
Before students can learn directly from symbolic representations like formulae, mathematics teachers must communicate mathematical ideas to them using natural language -- and with just a few iterations of correcting misunderstandings, this process somehow converges on the students having the same understanding of these abstract ideas.
That is, natural language succeeds here in bootstrapping a more precise form of communication.
Though for law, I think some ambiguity is beneficial. We should be going after the intent of the law, not the letter. This isn't just about bad encoding, as in not well aligned with the intent, but that there's always exceptions. Having that human judge be there to determine if something is actually reasonable or not is beneficial, even if there's a strong bias to follow the letter.
Ambiguity means that there are two or more possible interpretations and it's not clear which of them is intended. That's hardly useful. What's beneficial, and what you perhaps had in mind, is some amount of under-specification where the meaning is clear but leaves gaps to be filled in by judges.
On the other side there may be one person that wants to have sex with prostitutes himself, another that believes women should be able to do what they want with their bodies, a third that believes prostitutes can be an important way for young men to gain sexual experience and skill, a fourth that thinks prostitution is bad but legalization to be a way of harm reduction.
Not all of these people may be willing to admit their reasoning in writing. You could say that only following the written down reasoning is a feature. I haven't thought a lot about that subject, so I haven't made up my mind on it.
I want my politicians speaking more honestly. Or at least having to write things on the record. Things in the system that pressure more honesty, accountability, and transparency are better.
They are not willing to admit it in any domain, which is exactly why we want it in writing.
It's called accountability.
What I would like to see is what harm it is supposed to prevent.
What's more, the very "loophole" that enables a murderer to get away with murder is what an innocent person would use to be unjustly convicted of murder.
There are very few rules that will only permit good outcomes and never bad ones.
Incidentally, I think that the reasons why programmers tend to do that are quite similar to the reasons for using legalese that the paper identifies.
Take the phrase "Open Source" as an example. Us old folk ascribe specific meaning to that term - typically based on the legalese in Open Source licenses.
However the next generation have imbued it with their own (various) definitions. This leads to endless back and forth. For example I recently pointed out that SQLite is Public Donain, not Open Source. (With predictable pushback.) Today, in other thread someone claimed "its not really open Source unless its in git, and on github".
And the distinction between Free Software and Open Source is seldom understood.
So yeah, legal documents are gard to parse because they can't take "common meaning" for granted.
Heh, mine is even more strict: to me, it's not really open source unless I can build it since if I cannot compile the project, I cannot change it for my needs and/or send those tested changes back upstream
I have a second 2nd level "requirement" about packaging it in a sane distribution format, because I don't think any reasonable person wants to have a .desktop file that is $(cd /home/src/foo; npx run whatever "$@"). I'm looking at you, Chromium, since I can get it to build just fine but count the number of hand-rolled /usr/bin/install calls https://github.com/archlinuxarm/PKGBUILDs/blob/741f8edf84c7b... because evidently the $(make DESTDIR= install) is just kidding
Work on big technical projects like Linux was also a strong signal for employers. For years now that signal has been a target to emulate so a. Lot of “open source” became FAANG resume building.
I get where you are coming from though. The choice of the word "Free" in "Free Software" isn't ideal because the word "Free" has multiple meanings. Most people (especially non-tech users) assume it means "no cost" rather than "unencumbered".
So yes, it would have been simpler either a different name. Hindsight is perfect.
That said Free Software is different to Open Source (although lots of tech folk conflate the two.) That aside, it doesn't stop people adding their own (incorrect) requirements or expectations, as seen elsewhere in this thread.
Legalese looks an awful lot like code (I'm a programmer) meant to cover all edge cases and not permitting subroutines. It would be an awful lot clearer if they used a single term and then defined it after the main body of the law. Put as much as possible into a global appendix--as a layman occasionally trying to look up a law I find eternal references to as defined by xxxx.
And require the state to publish a version with hyperlinks and hover text.
When someone says: "That's so random", it isn't a commentary on determinism. There are many cases where adhering to a precise definition becomes problematic in popular discourse.