English Language & Usage: Multi-Layered Discourse Room

Kit Z. Fox

1:06 AM

One of the cats has parked herself directly outside my eldest's door, for not discernible reason.

Mitch

1:18 AM

@KitZ.Fox No, the cat knows.

user174558

Is it a female cat? Maybe it is falling in love with your son.

user174558

@Mitch There was a little drama in the math room. I think I won't go there anymore.

Kit Z. Fox

@Mitch maybe she just wants to lie in front of doors.

user174558

@Mitch Do you know much about number theory?

Mitch

@JasperLoy ooh. drama. Like 'The Cherry Orchard' or 'She Stoops to Conquer'

@KitZ.Fox maybe she likes the draft? or she's just trying out that spot randomly to mess with you.

user174558

1:28 AM

@Mitch I have already divined the reason.

Mitch

@JasperLoy I don't get quadratic reciprocity. I just don't get the point.

Kit Z. Fox

@Mitch Oh that's a thought.

Mitch

But I really like linear diophantine equations.

user174558

@Mitch I just wanted to ask if you like any number theory book.

Mitch

@JasperLoy well, duh, hardy and wright

user174558

1:30 AM

@Mitch Too ancient for me. I found two covering elementary, algebraic and analytic all in one.

Mitch

which ones?

user174558

Ireland, Rosen: Classical Introduction to Modern Number Theory

user174558

Baker: Comprehensive Course in Number Theory

tchrist

@Mitch chinmaylokesh.wordpress.com/2012/01/27/…

But I have that in a book I wrote somewhere.

It's not new. Doug McIlroy showed it long ago.

user174558

Who knew math is used in programming?

user174558

1:35 AM

@terdon and @AndrewLeach, and non-authors simply eat lunch.

tchrist

People are lined up your door. There are a lot of them. You must pick one person at random. Each person has the same chance as the next person. There is only room for three people in your room at the same time. How do you pick that person fairly when you don't know how many there are until you run out?

This is actually the problem of choosing a random line from a file of unknown length.

Most people read the entire file into memory, counting the lines as they come in, then pick a random one.

That won't work if you have 50 trillion lines.

Or 50 trillion people.

The chance of choosing a person at random is 1/N for the Nth person. Do that each time as N increases, and it will be fair.

When you pick someone, move them aside and have the next person come in. N has gone up.

The first person always has to wait for the second, as you see.

The second person has a 1/2 chance of being picked.

If so, person 1 leaves and person 2 takes his spot.

Person 3 comes in. There is a 1/3 chance of picking him, and a 2/3 chance of not. And if you didn't, then those 2/3 are 50/50 for the other two people.

So that is fair.

Forever.

And you never had to keep more than one person standing waiting.

user174558

QED.

tchrist

It is the same with picking a random line from a stream of indeterminate length.

user174558

I am going to bed. poof

Kit Z. Fox

@tchrist Is that like picking your nose?

tchrist

1:42 AM

@KitZ.Fox No, because there's always somewhere to put the boogers.

I trot that one out when people say programming doesn't involve mathematics.

Since it requires a proof, and therefore you are creating a true mathematical thingamabob, not just adding up numbers.

Kit Z. Fox

Only people say that are the ones who have never done it.

tchrist

Eliminating race conditions feels like proving things sometimes.

You know, this is a funny sentence: "And you never had to keep more than one person standing waiting."

Mitch

@tchrist clever. but basically by backtracking attempts all possibilities. nice short tool.

tchrist

Be that as it may.

Mitch

@tchrist I don't get how people think, or want to think, that they are different. programming is mathematical thinking, just not with numbers usually.

Oxinabox

1:59 AM

I had to shuffle 90Gb of text line by line a few months back (from 10,000 files). I ended up deciding that being smart would take too long, and so loaded it only a machine with 45Gb of Ram, and gave it a lot of Swap. It chugged pretty bad.
The smart thing to do would have been to partition it (by file) into two 45Gb subsets, shuffled those produceing 5,000 new files each, and then repeated the shuffle a few times.

Anyways,

tchrist

Don't shuffle lines, shuffle indices.

Kit Z. Fox

@tchrist I had a question like that I was going to ask.

Oxinabox

Of course, that is how it was implemented, but if you wanted shuffled files on disk, you do need to use those indexes to select lines and rewrite the files to disk

(Shuffleing the indexes took no time at all, reading out the lines and rewriting to disk took several days).

---
Does English.SE not take [tag:reference-request]s? There is no tag for it, where as there is on other SE sites.

Kit Z. Fox

We do not.

You can ask in here though.

And there are some Meta questions with resources.

tchrist

Some SE sites take resource recommendations, but many do not.

Oxinabox

2:04 AM

Not resources, specific citable references.

tchrist

You mean when was this word first printed type references?

Or do you mean what's a good book to figure that out using?

Oxinabox

The former.

tchrist

Oh, ok then. Those can be turned into real questions if you just put some work into them.

Oxinabox

Roughly I have a theory that is given a particular (multi)set of words without knowledge of ordering, then one sentence (with ordering) which you can write using those words is normally much much more likely than the others.

tchrist

We don't have an especial tag for them that I can think of.

Oxinabox

2:07 AM

Its not always true, eg Show me flights from Paris to New York." is just as reasonable as "Show me flights from New York to Paris". On the other hand often one ordering is far more likely and reasonable than the others, "The dog chased the cat." is much more often correct than "The cat chased the dog.", it is also preferable over "The cat the dog chased." though all three are grammatical.

tchrist

I have bobcats and mountain lions hanging around. They can chase dogs if they feel like it. :)

Oxinabox

Yes, for sure it does occur, but that might be 1 time in every 10,000 uses of that set of words?

tchrist

Your "The cat the dog chased" example of embedding is infinitely extensible, but people's brains will melt down pretty quickly.

Oxinabox

tchrist I know, my highest voted question is on that. People brains do melt. This question isn't about center embedding though, it is just about given the words (without option to repeat) order them

So I am look for a paper that confirms (or denies) my theory based on corpus statistics. (so far the best i have is http://lsa.colorado.edu/papers/cogsci97.pdf, but I don't really like that).
My own results on the Brown corpus seem to agree with my theory. But I'ld rather cite a well known work on the topic.

tchrist

You might do better on Linguistics. They tend to have the researchy NLP stuff there.

Mitch

2:11 AM

@Oxinabox are you writing a paper? It may turn out that your paper is the zero-day paper.

Oxinabox

I am writing a paper, but I really hope mine isn't, I'm not an English major.

Mitch

Yes, Linguistics. Especially since there's nothing special about English.

Oxinabox

It might be special about English though.

tchrist

Aw gee, English is the most especialist language never doncha know.

Oxinabox

fair point

Mitch

2:12 AM

@Oxinabox the subject you're asking about is computational lingusitics. English majors just study the meaning in literature. rarely language things.

tchrist

sighs heavily

Oxinabox

You are correct Mitch. My point was I'm no linguist, English or otherwise.

Mitch

Oh. I took English major to mean "an English major in an English speaking university" which implies it would just be literature.

where are you looking for references?

you may want to check out text analytics, probabilistic grammars, ngrams

Oxinabox

Yes. indeed that is what I am working on.,

Basically it is Given a Bag of Words, how often can a good stochastic language model order them correctly?

Mitch

so I'm sure there are lots of papers that discuss it. it may even be so common as to be in text books.

Manning and ... some other dude Statistical language processing?

Oxinabox

2:18 AM

I just reach over and grabbed Manning and Schutze from the other side of my desk.

(as you were typing that)

Mitch

nlp.stanford.edu/fsnlp

ha ha. yes. that one

Markov models?

Oxinabox

Yeah, I'm solving it with Markov models. But I want a ground truth from outside as to how often I can hope to succeed

Nope, nothing concrete in Manning and Schutze. I suspect I checked in there a few weeks ago.

Mitch

@Oxinabox what about references in that chapter?

Oxinabox

No, it doesn't touch on order beyond stating that words have order, before going into composition syntax

Kit Z. Fox

2:35 AM

Is Kosmo still around anywhere?

He might know right?

Mitch

@Oxinabox ignoring syntax, ngrams pretty much says it all (if I get the direction of what you're doing)

Kit Z. Fox

Wow. Almost two years since he's been in chat.

Mitch

Didn't Kosmo leave in a huff?

Kit Z. Fox

He came back after a year or two. As a proud dad.

But it's true we didn't see much of him after the July incident.

July 2011?

I think that was the first flag incident I was around for.

Might have been the first for EL&U.

Oxinabox

0

Q: How often can the words in a sentence be rearranged to form different but similarly likely setentence

I have a conjecture that given a particular (multi)set of words without knowledge of ordering, then one ordering is normally much more likely than any others. Its not always true, Show me flights from Paris to New York." is just as reasonable as "Show me flights from New York to Paris"; and j...

Kit Z. Fox

2:44 AM

Or was that a flag incident? It wasn't, was it? That was actually over a question closure, wasn't it?

Mitch

2:54 AM

I think it was Jeff somehow closed it

Autocratically

Oxinabox

I suspect looking for Shannon Entropy of sentence word order might do it.

Mitch

Mm I don't think so. Shannon entropy is notoriously difficult to compute

Oxinabox

More, adding the key word shannon to "entropy pf setence word order" decreases the number of results about maximum entropy models.

/me needs to go look up the difference between shannon entropy and normal entropy

Kit Z. Fox

@Mitch Yeah, a question about a joke.

Mitch

Ha ha. Stupid jokes.

Kit Z. Fox

3:01 AM

Well, there was a huge thing in chat about it. Hardly surprising that I would forget the details. Seems like we have one about every six months or so. We should plot it on a timeline.

Oxinabox

(Shannon entropy is entropy. ok. Did not know it was hard to compute)

Mitch

Psychometric study

My mistake then. Oops. I wa thinking Shannon capacity which is definitely hard to compute

Shannon entropy is ok

user174558

9:28 AM

@dam There is another Jasper in ELL right? Seems you are always talking to him there.

user174558

10:11 AM

Hello @matt.

Matt E. Эллен

hi

Damkerng T.

10:24 AM

@JasperLoy There is another Jasper indeed, though he spends most of his time on the main site, rather than the chat.

Hi, @MattE.Эллен!

Matt E. Эллен

hi

user174558

@DamkerngT. Is your country stable now?

Damkerng T.

It's as stable as it can get.

tchrist

@DamkerngT. Oh I hope not. The Roman Republic is stable. :)

Damkerng T.

Hehe!

Actually, I'm more concerned about the world. I'm not sure what's really going on.

tchrist

10:40 AM

Is one ever?

user174558

I am not sure what is really going on in my mind.

Damkerng T.

@JasperLoy That reminds me of the Tripitaka you were reading last year. :-)

user174558

@DamkerngT. Well, I am not sure which directions my faith will take. Sometimes, I have the feeling that even Theravada is false.

Damkerng T.

Whatever that can bring peace to your mind is good. If Theravada doesn't seem to work for you now, you can come back to it any time later.

user174558

@DamkerngT. I no longer subscribe to Mahayana or Vajrayana. I have a feeling they are a perversion of the original teachings.

Damkerng T.

10:48 AM

I bought a translation of a part of it (by Jack Kornfield and Gil Fronsdal). I planned to read it on my Kindle. The funny thing is my Kindle reader was broken just a few weeks after I bought the book! (Still waiting for a new Kindle here.)

user174558

@DamkerngT. I avoid Kindle because often it does not display correctly, according to reviews on Amazon, especially when math or tables are involved.

Damkerng T.

Oh, Kindle doesn't really work when it's a math book or even a linguistic book!

user174558

And then who knows what it will not work for.

Damkerng T.

I think I returned one book because of the formatting.

tchrist

@JasperLoy Me.

user174558

10:51 AM

@tchrist Ambiguous!

tchrist

@JasperLoy Moi?

user174558

You know what it will not work for, or it does not work for you.

tchrist

ponders being ambigamous

user174558

Maybe I was wrong there.

tchrist

@JasperLoy Do not think it were not deliberate.

user174558

10:53 AM

I am bad at the subjunctive. I avoid it like the plague.

Damkerng T.

Wait, how would you pronounce Segway?

I think I just heard it pronounced as "swag-way" on Ellen.

user174558

Maybe a pun?

user174558

I would read it as Seg plus Way.

Damkerng T.

Probably, I didn't keep my eye on TV. :P

tchrist

@JasperLoy I think everyone does.

@JasperLoy My use of it there dates from Middle English, and probably stopped around then, too.

Damkerng T.

10:56 AM

It wasn't a real Segway, anyway. It was a mini-segway. Maybe that's why.

user174558

@tchrist Ah, and I am not even familiar with Modern English!

Matt E. Эллен

Segue into something

tchrist

Exactly. The segue word made me think Romantically. I was thinking in English with a bit of Romance ~~superimposed~~ overlaid, which is what Middle English in some senses was. In some Romance languages it would be normal to use the past subjunctive in "Do not imagine that it were blah", but fortunately we don't do that now in English.

Matt E. Эллен

always simplifying by increasing ambiguity

tchrist

Saying more by saying less.

0

Q: Does a verb need to be preceded by "to"?

A (very) common verb is "to be", another is "to have". But you can also say that "have" is a common verb. The question is, when does a verb (on its own) have to be preceded by the preposition "to"? Is it stylistic? Is there a rule?

verbs prepositions

I know what he’s really asking, but this is hardly clear in the question.

> Come and go are verbs of motion.

> To come and to go are verbs of motion.

He's talking about mentions.

But that's not in the question.

user174558

11:27 AM

I was in that now deleted room where Rathony insisted that it should be 'go to find that' instead of 'go find that', LOL.

user174558

I am sorry for gossipping. A little is fine.

user174558

@matt I have not shaved for two weeks. I now have a significant beard and moustache.

Matt E. Эллен

good for you

tchrist

11:43 AM

Another beard myth bites the dust.

Matt E. Эллен

1:46 PM

hmmm. Drupal thinks 123Aa; is a strong password. I'm not so sure

terdon

@MattE.Эллен Digits, upper and lower case letters and a non-alphanumeric character. What else could you possibly want? :P

Matt E. Эллен

2:12 PM

@terdon the classification to be lowered from "strong" to "not easily guessable, but still brute-force-able"

Mitch

@MattE.Эллен I think it's pretty safe. I use it all the time.

Um...

This delete button isn't working

Matt E. Эллен

don't worry, noöne else will read that

Mitch

Ha ha I know. No one cares

user116848

2:28 PM

Hi!

Kit Z. Fox

Hi @Arrowfar.

user116848

Howdy @KitZ.Fox!

user116848

I have a vocabulary question.

user116848

We sometimes say xyz discussion is "unproductive". So can we use the terms "non productive", "unproductive" and "counter productive" interchangeably in this sense?

Kit Z. Fox

Hmm.

I'll think about it. They mean slightly different things. I've got a meeting though.

bbl

user116848

2:38 PM

See ya!

Mr. Shiny and New 安宇

@Arrowfar unproductive sounds more natural to me than non-productive and I think they mean the same thing: nothing was produced. Counter-productive is stronger: it impedes or reduces production. So a counter-productive meeting harms the team's progress.

whereas an unproductive meeting just doesn't improve the progress.

user116848

@Mr.ShinyandNew安宇 Ah I see. Thanks!

Mr. Shiny and New 安宇

np

Mitch

3:21 PM

Fargo is interesting, but why do so many people have to die.

It strains ... uh ... credibility? credulity? credulousness? verisimilitude?

They have a 'disclaimer' at the beginning of every show saying 'This is a true story. The names have been changed..blah blah blh'

Kit Z. Fox

@Arrowfar what Mr.Shiny said.

Mitch

But it's not a true story.

In the story, it's a true story. But in real life, it's not.

Why don't you just say Beowulf was a true story. I mean, it could have happened. Well, maybe Grendel and his mom but not the whale. You'd die of hypothermia or drowning before you could kill a whale. I'm just sayin.

Matt E. Эллен

3:53 PM

@Mitch I didn't. true story. (in the story)

Kit Z. Fox

4:11 PM

3 more hours until the SO election yields results!

terdon

drumrolllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllll‌lllllllllllllllllllll

Cerberus

5:47 PM

in The Incomprehensible Room, 38 mins ago, by RegDwigнt

@KitZ.Fox I said, "if I come in here and you call me a fag, that's one thing; but if I come in here and you are talking about kittehs, and I specifically search the transcript for mentions of fags, that's a whole nother story, and if I'm offended at specifically finding what I'm specifically looking for, that's my own fault".

user174558

@KitZ.Fox Are you running for mod there? =)

user174558

Is it true that Rob and Reg are not coming to this room anymore?

Matt E. Эллен

@terdon noöne's going to drum roll for 3 hours, surely?

terdon

They can try.

Matt E. Эллен

@JasperLoy depends what you mean by anymore. I saw Reg yesterday I think. Robusto has decided to take a break for a while and see how he feels.

user174558

5:54 PM

Jon Skeet should run for mod and not answer any questions, just do mod work all day.

user174558

I have many mouse freezes using Cinnamon desktop, should change to MATE desktop.

Transcript for

English Language & Usage: Multi-Layer…