A few words about that ten-million-dollar serial comma (2017) | Hacker News Reader

A few words about that ten-million-dollar serial comma (2017)

newyorker.com

69 pointshoffmannesque90 comments

http://web.archive.org/web/20210824190938/https://www.newyor...

> “The next day, I enjoyed pan-roasted oysters with a tomato sauce over rice, broccoli salad and bread pudding with chocolate sauce.” A comma after “broccoli salad” would have cleared the table before dessert.

I think if you write a sentence that's highly dependent on the reader paying careful attention to where an easy-to-miss comma is, you assume all readers know what Oxford commas are, and the sentence generates a lot of discussion about how it can be misinterpreted, it's just a bad sentence. I think instead of debating about Oxford commas as a band aid, you should restructure the sentence so most people will understand it first time without having to think too hard.

This feels similar to overly clever code which goes against the principle of least surprise that shouldn't pass code review.

“The next day, I enjoyed pan-roasted oysters with a tomato sauce over rice and broccoli salad, followed by bread pudding with chocolate sauce.”

Edit: Fair point that the dessert is clear now but not the main. Either way, I'd prefer restructuring the sentence to make it clearer where practical e.g. adding in a few extra words like "with a side of", "dessert", "followed by".

This is likely just me, but I find sentences where the independent clause is bookended by an introductory and dependent clause to be inelegant. I never write them.

It's easily noticed when omitted improperly and more easily overlooked when done correctly.

Can you give an example?

I'm a fan of hyphens:

"...oysters with a tomato sauce over rice, broccoli salad - and bread pudding ..."

I'm having palpitations at this.

I'm a big fan of em dashes, which—when used like this—signal a parenthetical remark.

Parentheses are way better than em-dashes for parenthetical remarks.

A set of parentheses make the syntax unambiguous.

An em-dash is “that multifaceted, all-purpose punctuation mark that can act as a parenthesis, comma, colon, semicolon or even quotation mark”.

Also I have spent too much time with computer languages, and I personally find the lack of spacing around the parenthetical em-dash to be an ugly syntax. Then again, I don’t closely adhere to the English rules of punctuation, because some of the standard rules make for ugly writing IMHO.

Mostly I try to avoid parenthetical statements because they are often a sign of lazy wishy-washy writing. I guess em-dashes are OK in novels where the text context usually makes the semantic meaning clear enough.

Edit: “The Chicago Manual as a style guide calls for no spacing before and after dashes. In British usage and in some style guides, spaced en dashes are used place of an em dash. Sometimes the space you see before and after the dash is actually a special character called a ‘hair space’ or a ‘thin space.’”. https://danieljtortora.com/blog/en-dashes-and-em-dashes-part...

Edit: “Most newspapers — and all that follow AP style — insert a space before and after the em dash.”. https://www.thepunctuationguide.com/em-dash.html

Edit: From C2: “U+2014 EM DASH is used to make a break—like this—in the flow of a sentence. (Some typographers prefer to use U+2013 EN DASH set off with spaces – like this – to make the same kind of break.) This kind of dash is commonly represented with a typewriter as a double-hyphen.”

Edit: Also see: https://en.wikipedia.org/wiki/Dash#Em_dash

Functionally, I guess the only advantage that the em-dash has is that it fails loudly and miserably when nested. For example:

Parentheticals -- already a bad -- although sometimes unavoidable -- habit of mine -- should never be nested.

Completely incomprehensible.

Parentheticals (already a bad (although sometimes unavoidable) habit of mine) should never be nested.

Dangerously comprehensible.

However, em-dashes do make the reading voice in my head change over to an old-timey radio announcer so that is cool.

> However, em-dashes do make the reading voice in my head change over to an old-timey radio announcer so that is cool.

That is cool.

When I read a parenthetical in parentheses it's like someone is talking to me while I'm trying to watch a movie: Even if the information is useful, it's annoyingly distracting. The same parenthetical in em-dash form doesn't have this same effect on me.

However—and I hope you'll bear with me here—for the most part the information in a parenthetical is either not needed or important enough to justify its own complete sentence.

I am a writer, and fully conversant with the fact that an em dash signals a break in the flow of a sentence, and one such use of a break in the flow is to use two of them to introduce a remark about the sentence itself, and then return to the sentence.

If you think parentheses are "way better," use them as you see fit. I can't imagine using both in the same sentence, but both are in my toolbox and I use whichever one suits the tone and content I'm communicating. (I often prefer parentheses for complete sentences.)

As for you spending too much time with computer languages and not liking the lack of spacing around em dashes, I believe you but don't share your taste at all. What am I to make of 1+(2*3)? Why do we not need spaces around operators and parentheses in arithmetic, but somehow it's ugly in prose?

Why are you ok with no space between the last word of this sentence and the interrogative? If you don't like a lack of spaces around em dashes, what do you think about "quotes?"

I am extremely familiar with en dashes and thin spaces. When writing text intended to be set with variable width, I put thin spaces on either side of en dashes but not em dashes. So I use thin spaces around an en dash 3 - 4 times a day.

What you have is a kind of taste. It's not universal. But it's yours, and if it serves you and your readers, it's fine.

Aren't semicolons the correct english list separator?

That's another argument in itself. I was always taught that phrases set apart by semicolons need to be capable of standing alone as complete sentences, so it's jarring to see them used to delimit nouns or noun phrases in a list.

At the same time, the language clearly needs a comma-like operator with higher precedence than the comma, and it's very hard to argue against pressing the semicolon into that role.

The semicolon is often used as a superlist separator, with commas used for the sublists.

Yes, that's the point -- it's often used that way, but not "correctly."

What's incorrect about it?

As I said, at one time it was taught that clauses separated by semicolons should have their own subjects and verbs.

> At the same time, the language clearly needs a comma-like operator with higher precedence than the comma,

When a bulleted list is too much, I usually fall back on 1) this thing 2) this, this and this other thing and 3) this final thing. Gives you an easy way to refer back to each item too. Bit stuffy for causal writing though.

> “The next day, I enjoyed pan-roasted oysters with a tomato sauce over rice and broccoli salad, followed by bread pudding with chocolate sauce.”

This just moves the ambiguity.

Is the tomato sauce over rice, with a side of broccoli salad? Or is the sauce over [rice and broccoli salad]?

In this case "rice and broccoli salad" doesn't really sound like a dish, but why the fuck wouldn't you just use a tiny piece of syntax to make it super clear, rather than relying on the reader hopefully having similar culinary sensibilities?

How does a comma alone fix the sentence I quoted though? It could then be read as the tomato sauce going over the bread pudding. I think if you really want super clear, restructure the sentence further.

>“The next day, I enjoyed pan-roasted oysters with a tomato sauce over rice and broccoli salad, followed by bread pudding with chocolate sauce.”

Your version is even more ambiguous:

Were the oysters separate from or included in the tomato sauce, and was said sauce served over rice, or was it served on the broccoli salad as well? Or perhaps it was a salad comprised of both rice and broccoli, over which a tomato sauce was poured - either with or without the pan seared oysters?

Just use the freaking commas - it makes everyone's life a lot easier.

Wat this all is, to my eyes, is people putting form over function. If you care about the message, you consider the people receiving it and try to reduce ambiguity as much as possible. If you care about the medium, then you make trade-offs about aesthetics which may reduce legibility but appeal to some other quality that is being considered.

There's obviously a place for both. Poetry does not necessarily benefit by removing ambiguity. Technical manuals do. Depending on what the author is attempting to do, different choices may be made.

In contracts meant to define actions and behaviors, ambiguity is not beneficial (at least not for anyone that's not trying to swindle someone else). There are constructs or forms that can be used to reduce ambiguity, and they should be used where appropriate. For example, if a bulleted list is allowed, that removes quite a bit of ambiguity in some cases.

Unfortunately, for contracts we are forced to use an imprecise medium (natural language) to define what we want to be a precise set of constraints. There will always be some friction there where one for not accurately account for what's meant by the other. Additionally, accurately describing all relevant facets of reality that might apply is hard.

The next day I enjoyed

  - pan-roasted oysters with a tomato sauce over rice
  - boccoli salad
  - bread pudding with chocolate sauce.

There, I fixed it.

There's still ambiguity as to whether it's:

   (pan-roasted oysters) with (a tomato sauce over rice)

or:

   (pan-roasted oysters with a tomato sauce) over rice

I don't think there is. It is the second. The clear convention is that each meal item is a list element.

It's not clear to me.

> (pan-roasted oysters with a tomato sauce) over rice

I picture a bowl of rice, then oysters on top of the rice, and then tomato sauce poured on top of the oysters.

> (pan-roasted oysters) with (a tomato sauce over rice)

I picture a plate. On one half is oysters, the other half is rice with tomato sauce poured on it.

As someone who would never put tomato sauce on either oysters or rice, either are equally likely.

No, it's a list of two subsets contained in the set "meal."

     - main-dishes (x1)

     - side-dishes (x2)

The entree set has 1 item in it, the "pan-roasted oysters with a tomato sauce over rice.The ingredients of "entree" follow the same hierarchical structure as the ingredients of "meal," so:

     -main-ingredients (x1)

     -side-ingredients (x2)

Since we're only talking about two sets here, a comma separating them is unnecessary. You could use one if you wanted, ("pan-roasted oysters, with a tomato sauce over rice") but then the syntax of the sentence would be off. You'd have a list of three things:

     -pan roasted oysters

     -a tomato sauce over rice

     -broccoli salad and bread pudding with chocolate sauce

This is logically inconsistent because the oysters, tomatoes, and rice comprise a single dish. The tomatoes and rice don't stand on their own but describe how the oysters are prepared and served, just as the non-entrees describe how the meal is served.

Lets compare:

ORIGINAL: >“The next day, I enjoyed pan-roasted oysters with a tomato sauce over rice, broccoli salad and bread pudding with chocolate sauce."

OXFORD COMMA: > “The next day, I enjoyed pan-roasted oysters, with a tomato sauce over rice, broccoli salad, and bread pudding with chocolate sauce."

By structuring the sentence in that way, you erase any information about the value of the meal, and the relationships between its components. In return you ensure that someone won't think there was chocolate sauce on the broccoli salad.

I would write "... and bread pudding with chocolate sauce as desert."

If a reader didn't know what bread pudding was, they could conclude that you had just chocolate sauce for your desert.

Trying to write English like you write code is pointless. It's edge cases all the way down and there aren't any shortcuts for gathering them all up except to assume a certain level of intelligence in your reader. Writers should focus on elegant, easy-to-read text that expresses their thoughts and the relevant details in a manner their main audience can understand.

I agree with the top-level commenter that the Oxford comma is used where a rephrasing may be better. Of course, most phrases can be better written, so it's largely meaningless. The Oxford comma is concise and unreasonably effective. It eases reading and should be adopted universally.

My two cents against the use of the Oxford comma would be that all of Europe's languages are written without it. Surely if neither German, nor Spanish, nor Finnish, nor Lithuanian, nor Hungarian, nor Bulgarian - if none of them need the Oxford comma, English probably doesn't either.

Without regard to the actual conclusion, I don't one should necessarily apply lessons learned in sensible languages to this abominable offspring of German and French that we, for some reason, use.

> “The next day, I enjoyed pan-roasted oysters, with a tomato, sauce over rice, and broccoli, salad, followed by bread, pudding, with chocolate sauce.”

yum ;P

> I think if you write a sentence that's highly dependent on the reader paying careful attention to where an easy-to-miss comma is, ... it's just a bad sentence.

Consider the function of the comma in "Time to eat, Grandma" vs "Time to eat Grandma".

Proper punctuation saves lives.

The commas in the second Amendment and their meaning are probably worth Trillions of dollars. When you think of all the private arms and all the costs of gun violence in the US over the years which hinge on the individual right to bear arms, when punctuation in the second amendment can lead people to read the text as only holding a right to bear arms within a well-regulated Militia. Turns out that the commas weren’t even standardized and lots of differently punctuated versions were ratified by the different states.

https://en.m.wikipedia.org/wiki/Second_Amendment_to_the_Unit...

I think that people would have interpreted that amendment the way they wanted to regardless of comma placement. They already blind themselves to the "well regulated militia" wording entirely, so ignoring an extra comma or two makes little difference.

Blind to and blinded by, depending on who you ask.

The meaning seems obvious. A well regulated militia is required for the security of a free state. Therefore, people's freedom to keep and bear arms shall not be infringed. It declares an already existing freedom to keep and bear arms (as IIRC nothing on the federal level forbade it), and forbids its infringement by the state, with the goal of keeping the population familiar with the operation of firearms for the benefit of local militias.

Or read the longer explanation by the Cornell Law School https://www.law.cornell.edu/constitution-conan/amendment-2/s... .

That explanation you linked is a whole lot of words, and after reading it I am less sure of what I thought the Second Amendment means than ever. Its meaning is anything but obvious.

I say this because there is nothing about self-defense in the text of the Second Amendment, and yet the SCOTUS has held that self-defense is one of the rights that cannot be infringed. It's not at all obvious how a statement about the necessity of a well-regulated militia has anything to do with personal self-defense with a handgun. I can see a world where the DC handgun ban was upheld by the SCOTUS because long guns (for the well-regulated militia) were still available and personal self-defense isn't covered under the wording of 2A. Of course, if you bring the zeitgeist into view this seems absurd, but in a vacuum where the 2A is the only reference it makes logical sense.

Well, we are talking about a body that finds the constitution protects privacy, even though CTRL-F 'privacy' has zero results.

I am not saying they are right or wrong, but any pretense that the courts rule based on the text of the law should be discarded.

The reality is they find rights and laws that seem reasonable to their own selves, and disregard precedent that is poorly argued (ie they don't like the result).

The alternative is perhaps that we are stuck with Plessy v Ferguson until the end of time or a constitutional amendment is ratified (end of time seems more plausible).

I'd argue the 4th Amendment gives you a right to privacy from the government.

Textualists might be able to make your argument. Any constitutional scholar worth their salt could not.

The only way to misread the Second Amendment is to assume that it, alone among the other amendments in the original Bill of Rights, was written to constrain the people rather than the government.

In other words, you have to approach the question with an agenda -- and a conclusion -- already in mind. Bad for intellectual honesty regardless of how you feel about gun ownership rights.

Which government? Two states still had established churches when the First Amendment took effect.

If the framers hadn't felt the First Amendment to be necessary, based on their awareness of both classical and all-too-recent history, they wouldn't have bothered including it. Arguably the same is true for the Second Amendment.

They didn't – Madison and Hamilton were backed into a corner during the ratification process by the anti-Federalists.

(Shrug) What ended up in the document is what counts, the rest is just commentary.

This sort of sanity in interpreting written texts is disallowed on this forum.

Here, guns are bad. The second amendment was meant to take away guns from the people and allow them only for the government, specifically, the armed forces.

This is because here, we love tyranny, and are rather tired of the uppity flyover staters who are holding it back.

Who is reading the Second Amendment as constraining the people rather than the government?

It seems like it is either constraining the government from restricting the private ownership and use of firearms altogether, or it is constraining the government from restricting the private ownership and use of firearms for the purpose of maintaining a militia.

Who is reading the Second Amendment as constraining the people rather than the government?

Everyone who argues that the "militia" means the National Guard or another organized, government-sanctioned body.

While the US Constitution might not strictly define militia as government-sanctioned, nearly every (if not every) state does so either in state level Constitutions or legislation.

Private militias are illegal

> While the US Constitution might not strictly define militia as government-sanctioned

The US Constitution presupposes a generally universal militia and US statute law explicitly defines the federal militia as consisting of all able bodied males 17-45 who are, or have made a declaration of intent to become, US citizens, plus those over 45 in or eligible to join the National Guard, plus all females in the National Guard.

> Private militias are illegal

Private militias are not generally illegal, but they are generally prohibited by each state from engaging in actions reserved for the state militia.

There are examples where using an Oxford comma makes a sentence ambiguous and there are examples where not using an Oxford comma makes a sentence ambiguous.

In all the examples I've seen where using an Oxford comma makes a sentence ambiguous that sentence contains an appositive phrase that is set off with commas and that is what creates the ambiguity.

Commas are the most common way to set off an appositive phrase, but it is generally considered acceptable to use em dashes or parenthesis instead. Switch to using one of those instead of commas for your appositives, always use the Oxford comma, and I think that clears up all the ambiguous cases.

Why, commas not withstanding, should Maine prevent people in the food supply chain business from getting overtime. I am pinko-Euro-liberal and even I think that's government meddling. Is there a historic reason for this if just labour-bashing ?

"During the New Deal Era, President Roosevelt struck a bargain with Southern Democrats: they would support worker rights legislation so long as their farmworkers (and other predominantly African-American workers, such as domestic workers) were exempt. Thus, Congress excluded farmworkers from the National Labor Relations Act of 1935 (NLRA), the main federal law that protects workers who join and organize labor unions, and from the federal minimum wage and overtime protections in the Fair Labor Standards Act of 1938"

https://thenaturalfarmer.org/article/brief-history-farmworke...

The government isn't preventing them from negotiating a better contract with their employers, it's just refusing to enshrine their entitlement to overtime into state law. From that point of view, this carve-out reduces government interference in labour relations.

I think it could be argued that carving out an exception counts as increased interference because it means the government is inserting itself between all workers and employers in order to mandate job classifications. Without exceptions, businesses would not have to spend time/money defending their assertion that this worker is A and that worker is B and thus is less entitled to greater pay.

The governments here in the US fill two roles. Oppressing workers and protecting them. You never know what it is going to be from one moment to the next and we all await verdicts to cases like this, not knowing WTF is going to happen.

Unions are protected with various laws...then court cases pops up where there is some debate about if a Union can legally strike or even negotiate.

In my mind it is very clear. Do the workers have leverage? If so, then "the company??" has to negotiate...just like they would with any business partner.

In the end the courts are used as a bulwark against worker rights. Any victories for workers just give substance to the idea of governments intervening in business negotiations when board rooms want to socialize expenses.

Easy to miss line in the article says it’s due to avoiding spoilage in harvested food.

Absolutely labor bashing.

> Lest we lose perspective, this law on the books of the State of Maine applies to people who work with perishable foods, and the point is that pokey employees should not be rewarded for taking their sweet time getting the goods to market.

You could penalize that without penalizing those that work long hours for legitimate reasons.

> Why, commas not withstanding, should Maine prevent people in the food supply chain business from getting overtime.

This provision doesn't prohibit overtime pay, it excludes that group of people from government-mandated overtime pay. A company could still choose to pay it's workers overtime, or workers could demand/negotiate that.

personally, i generally find the absence of an oxford comma slightly jarring even without any ambiguity issue- i feel it matches the way i read and speak (midwest en-us)

conversely, i only use semicolons in javascript when required..

Same. Seeing text without them bothers me unreasonably. I know some writing guides advise not using them, but I reject that reality and substitute my own.

Missing caps on the first letter of a sentence feels kinda odd also.

I'm just seeing arrays everywhere. 1, 2, and 3 seems wrong to me. I reckon it should either be 1, 2, 3 or 1, 2 and 3. When a comma is followed directly by 'and' (or by 'or'), one of them just seems superfluous.

On a side note, it makes you wonder if lawyers should take a course in programming. I wonder if these two arrays would have clarified matters:

The Intent of the Contract: no overtime for the following: [ canning, processing, preserving, freezing, drying, marketing, storing, packing for shipment, distribution ] This would have included truck drivers because it separates packing from distribution.

The Ruling: No overtime for the following: [ canning, processing, preserving, freezing, drying, marketing, storing, packing for [ shipment, distribution ] ] This excludes truck drivers because it puts distribution with packing and truck drivers don't pack! The power of arrays.

It really does seem like simply adding more punctuation could save a lot of trouble.

Ahh, one of my pet peeves! I like and use the Oxford comma in my emails but have observed am in the extreme minority on this one.

Here’s an English SE question with some interesting answers: https://english.stackexchange.com/questions/412/should-i-put....

Here’s Another answer with great visual illustration of the strippers, JFK, and Stalin example: https://english.stackexchange.com/questions/172671/oxford-co....

My other sensibilities include not putting two spaces at sentence end (among other things this shows that you’re most probably 40+) and writing PowerPoint slide titles in sentence case rather than title case.

>My other sensibilities include not putting two spaces at sentence end (among other things this shows that you’re most probably 40+)

That's a silly generalization. I'm only 39.

> but have observed am in the extreme minority on this one.

The desire to make sentences unambiguous seems to be very rare, so this isn't too surprising.

I wonder if programmers use the Oxford comma more than the general population. It just feels more consistent to me to have all items delineated with commas (or semicolon if items contain commas: http://sana.aalto.fi/awe/punctuation/semicolons/02.html).

It's not consistent, though. Nobody uses the Oxford comma in a list of two.

In the Oxford regime, "apples and oranges", when prefixed with another fruit, becomes "bananas, apples, and oranges".

If you wanted to implement an "insert commas" algorithm recursively, choosing Oxford requires passing an extra bit along to signal to the terminal case (joining the last two list elements) that there are > 2 elements total. Inelegant to my mind.

If I say out lout "bananas, apples, and oranges", then I have a slight pause between each item in the list and the comma represents that pause.

I say "bananas (pause) apples (pause) and oranges". I do not say "bananas (pause) apples and oranges".

It definitely appeals the my sense of logic and clarity.

I don't like the Oxford comma, because I see a, b, c <-> a and b and c, i.e. the comma is a stand-in for the word "and". Thus "a, b, and c" has a double separator between b and c.

> My other sensibilities include not putting two spaces at sentence end (among other things this shows that you’re most probably 40+) and writing PowerPoint slide titles in sentence case rather than title case.

Yep, guilty on both counts, and over 40. I feel no particular inclination to change, either, since neither seems to be anything more than personal preference. Someone will always disagree, so I just stick to the way I've always done it.

The "two spaces" thing also complicated by the fact that in so many contexts the extra space is simply ignored, as browsers ignore additional whitespace by default and so much writing and reading takes place through a browser or some other thing rendering HTML these days.

Of course you can do things like "pre" formatting or insertion of non-breaking spaces to force their rendering, but that's somewhat rare. I wonder if HN does that? Doesn't look like it.

I find this muddied even more by the shortcut for "start a new sentence" being a double space on mobile keyboards. Seems like a habit I'd pick up even if I'd learned to type on a phone.

Many people use the Oxford comma in every case because they don’t trust their own judgement about ambiguity. That’s always struck me as pretty lazy, despite proponents insisting it’s some sort of principled stance.

A writer should strive for clarity over style whenever possible.

A poet sits in the dappled shade of a tree's summer embrace, face flashing momentarily in disagreement.

> That’s always struck me as pretty lazy, despite proponents insisting it’s some sort of principled stance.

You don't need the comma before "despite", but I don't think you're lazy because of that.

If you don't understand Oxford Commas, you aren't fluent in English.

Brooklynese has entered the chat, wearing its NY Times press badge.

There is a classic programming joke where, to my eyes, the Oxford comma is a must:

"There are two hard problems in computer science: naming things, cache invalidation, and off by one errors."

Without the comma it seems weird to me:

"There are two hard problems in computer science: naming things, cache invalidation and off by one errors."

"the State of Maine specifically instructs drafters of legal statutes not to use the serial comma"

Why would they ever suggest not using the comma? Maybe they'll change the instructions now.

And the $2M comma issue from Canada https://www.theglobeandmail.com/report-on-business/the-2-mil...

And Rogers Won.

http://www.slaw.ca/2007/08/22/rogers-wins-comma-contract-dis...

  Language buffs take note -- Page 7 of the contract states: The agreement "shall continue in force for a period of five years from the date it is made, and thereafter for successive five year terms, unless and until terminated by one year prior notice in writing by either party."
  Rogers' intent in 2002 was to lock into a long-term deal of at least five years. But when regulators with the Canadian Radio-television and Telecommunications Commission (CRTC) parsed the wording, they reached another conclusion.
  The validity of the contract and the millions of dollars at stake all came down to one point -- the second comma in the sentence.
  Had it not been there, the right to cancel wouldn't have applied to the first five years of the contract and Rogers would be protected from the higher rates it now faces.
  "Based on the rules of punctuation," the comma in question "allows for the termination of the [contract]at any time, without cause, upon one-year's written notice," the regulator said.

Honestly, that's just ambiguous in general. The one year prior notice should specify that it is prior to each renewal.

Isn't the trailing comma at the end of lists (like Python, JavaScript, C#, and this sentence,) just an extension of the serial comma by one more step?

Reply on news.ycombinator.com