Over on my board, we’ve been amusing ourselves with this toy. You enter a short text, and an AI will then attempt to continue it.

You may remember my Markov generator. This is similar, but uses a neural network. The really impressive thing is the ability to keep on topic! I assume it’s using parts of a corpus, but it’s able to continue using words you suggested.

Here are some of my attempts; my starter text is in bold.

When the aliens came, no one was quite prepared to see what came out of their gleaming blue spacecraft: ersatz humanoids, but in real human form. With no sense of the Earth, they had no reason to attack New York City.

The aliens returned six weeks later with the first prototype: a new, streamlined model for the ship that still looked rather human, but now glowed blue. A group of young scientists in a nearby laboratory, now working in harmony with the machines, began work on a “computer” that could communicate with the craft’s computer.

The most important element of this “computer” was the ability to speak alien languages. The ship quickly became a popular meeting point among alien visitors. In fact, a number of countries soon decided to take over the job of managing this new alien visitor.


Although the existence of God cannot be proven, scientists now believe that there is evidence for the existence of godlings. Although scientists agree that there is life on other planets, they do not believe there is life on planet Earth.

As Godling, Adam is seen in biblical accounts as being a being that had great knowledge, wisdom, skill, and courage.

Adam was known to have given his brother the serpent a rod, which was so powerful the earth trembled under it. And God was also given a man whose name was Seth, whom Adam called his brother, and he became father of all the living creatures on earth. (Genesis 3:1-2)


I have a killer idea for a video game. You can play as anyone you like from history, and replay their life in their own culture. Let’s call it “Nostalgia: It’s the way we feel when we were kids, or our parents were.”

You’ll be the first to hear about my new game if you help me raise $1,000 to pay for all the artists, people, marketing, and travel to Montreal. I want to get it out there so that we can show how we can make a game with real people, from people who didn’t expect it to take off like this. This game is going to make a lot of people’s families cry. As you might have read in the last post, “All the art I use is original. All the art is original.” All the art is for sale – it doesn’t exist anywhere else. But a lot of people seem to be confused about where to look for it.


In the year 2510, the remains of human civilization are run by three corporations: Facebook, Google, and Walmart.

Walmart is a corporation founded by a former United States Marine named Sam Walton. In the year 2510, after the government’s attempt to shut down Walmart has fallen flat, Walmart decides to shut down Walmart. One of the key differences between Walmart and most other corporations is: The majority of the population can’t survive without Walmart’s products or services. It is also a company that can sell its products or services even in other countries at a lower price than its competitors.

After a number of years, a series of disasters leads to Walmart’s demise. This leads to the “Walmart Effect”: After Walmart’s demise, a whole chain of stores, restaurants, department stores, hardware stores, and grocery stores are created, with the name “Walmart”. A number of other Walmart franchises and subsidiaries become available, often to other companies that are in conflict with Walmart. These corporations provide products or services to the populace that are more in tune with their worldview.


From ZBB member alice:

When I told the ZBB about my new theory which combined syntax, pickles, and tea, she tried to give me the run down she had about the ZBB being a place for people to read about my theories. I went to the ZBB as a new visitor a few times and didn’t even leave a note so I think they might have figured that.

I think the main reason why I didn’t go back to the ZBB is because I liked the layout. I didn’t like the idea that the only way I was going to get to read about my theories was to buy a book, a tablet, and send them via airmail. I wanted to learn about the ideas I had come up with, not just go to a bookstore.

I also like going to other coffee shops, but coffee isn’t the only thing I like to drink!

Raphael found this one:

It was twenty years ago today. Sergeant Pepper told the band to play. The band played, until they couldn’t anymore. It was the longest, most embarrassing performance at a music festival in history. They didn’t even start it, they just ended it.

In the twenty-three years since that day, Pepper has had to wear many uniforms, in many different countries. He has had to do a lot of traveling – to Europe, the Middle East, South Africa. He has had to make many compromises, to keep everything intact. For the past twenty-three years he has lived alone – in a large, modern brick apartment at the top of the same building (he could never bring himself to live alone). His apartment was as clean as a church’s. It was the sort of apartment you might take with you in a warzone, if you had to. He had gotten so involved in this band (he has since moved onto another project), that he had become a bit overwhelmed by the idea of not sharing it with a partner. He hated to share it with a partner, especially since he didn’t sleep very well.

And from Pabappa. Note that the AI continues to use the supplied name Repilia, but ‘decides’ that it’s not a country but a woman.

“Our war against Repilia must come to an end. Repilia will rule our homeland; our men will be disarmed. There shall be no future wars within our territory, as our battles were fought in disobedience of Gahanaya.” When the women in the assembly heard this, they began to worry about the soldiers camped out to the north of the temple. “My, that wasn’t bad. I would have thought they’d have been happy to see our leader come back. But what about the soldiers south of the temple? It seems Repilia may need them.” “My lord, the northern forces were not under our commands. They are my soldiers, and I am not going to give them arms. That may be a little harsh, but I hope it’ll appease the women at the assembly. After all, Repilia is not the same woman she looked like just the two of us.” “If that’s the case, then the army must fall. The army that led us to Repilia is all that’s left in northern territory.”

So, tonight I finally saw Black Panther, which you may have heard of. If you haven’t, I suggest you go see it; it’s pretty good. This is actually the first Marvel movie I’ve seen. I hear they have, what, half a dozen by now?

We’ll get to the actual people below, but they have to work hard not to be upstaged by the set and costume design, which are some of the best to be seen in any sf/fantasy film. E.g. the big reveal of Africa’s biggest city, Wakanda:


Whoops, that is Africa’s biggest city, but it’s the real city of  Lagos, Nigeria.  Here’s Wakanda:


So, this is a really cool shot, and it kind of ruins one of my jokes, comparing Wakanda to Numbani from Overwatch. This is far better done, not least because it isn’t just futuristic slabs as in every other movie and video game; it has interesting textures and seems anchored to the natural world. The street scenes are great too– it looks like a lively city that definitely doesn’t look American.

Still, I included the picture of Lagos for a reason– as a reminder of how mind-bogglingly large it is (the metropolitan area houses 21 million people, a little more than New York), and that the continent isn’t the basket case some people depict it as. (Nigeria’s per capita income is about the same as India, which today we think of, or should, as a rising power.)

The movie itself has a lot to say about oppression and unfairly divided wealth, especially as it relates to Black people, but its view of Africa outside Wakanda is uniformly negative. It’s the “Third World” that Wakanda hides itself as; the only scene set in non-Wakandan Africa is a human-trafficking operation. Not every movie can be everything, but in this area the movie is maybe a little too American.

Now, superhero movies are kind of forced to have a stereotyped and somewhat dumb structure. First you have to show that the superhero is awesome: they go and beat up normal mooks in amazing ways. But since 90 minutes of beating up mooks would get old, you have to have a supervillain, and the hero has to be beaten, and it’s hard not to make them look incompetent. Finally they get to be awesome again and the villain is decisively overcome.

This was a major problem in The Dark Knight, and Black Panther can’t quite escape it. Chadwick Boseman gets his early awesome scenes, but he also spends a lot of the movie looking kind of lost.

There’s also a special problem with the Black Panther character, which– to be honest– was created by a couple of white guys with pretty retro ideas about Africa, full of rhinos and kings and acacia trees. That is, he’s a superhero but also a traditional king His country is supposed to be wealthy and technologically advanced, yet also an absolute monarchy. (The main driver of the movie’s plot is that the king is chosen via a fight to the death.) The political contradiction was faced in the comics by Ta-Nehisi Coates, but he and Boseman both have problems humanizing the king– both T’challas are regal and austere and a little humorless.  On the other hand, that does give him a real character arc, and by the end of the movie he does have something to smile about.

Fortunately for T’challa and the movie, he’s also surrounded by badass women who don’t have to go through that act-two round of doubt and defeat. The standouts here are his sister Shuri (Leticia Wright) and his main general Okoye (Danai Gurira). I would gladly watch a movie centering on either of them. Okoye is beautiful to watch, making the superheroics look effortless. Shuri has great fight scenes too, but she’s also Wakanda’s Q, its scientific heart, and there’s nothing like her smirky smile when she’s carelessly explaining some tech she knows her listeners won’t understand.

The main antagonist, Erik Killmonger, is unusually good for a supervillain, because Ryan Coogler (director and co-writer) gives him an intelligent ideology and plan. (And at least at first, he’s more likeable than T’challa.) He wants to fight back— he wants to use Wakandan technology to take over the world and “run it right.” When he get a chance to confront the Wakandans, he asks them what they were doing when Africa was being carved up and millions of its people enslaved. No one answers, because they have no answer. They were protecting their little turf and that’s it.

Now, the dude apparently wants to use terrorism to create this empire– his plan consists of shipping out weapons, which he’s hoping will be used to kill a lot of people. So, that’s pretty bad. But he’s useful as a critique of Wakandan complacency, and an object lesson in why alpha-male combat might not be the best political system. And again, all this is way more sophisticated than most superhero stories, which are mostly about supercriminals with no relation to actual crime, and near-supernatural threats with no relation to actual global threats.

A few minor cavils:

  • Bits of the plot were obviously storyboarded, but not thought out. E.g. the operation in Busan (hi D.Va!) made no sense at all: the artifact wasn’t recovered, not enough operatives were sent, and Klaue was not secured.
  • T’challa asks his frenemy M’baku to safeguard his mother while the capital is held by Killmonger. Then, to push an alliance, he says Killmonger will come after M’baku. These statements don’t seem compatible…
  • “Hanuman”?  Yikes.

Even more than the set design, the costume design is consistently great. Okoye and the rest of the all-female royal bodyguard are especially striking in their red armor. The designer went to the trouble of creating designs for each of Wakanda’s five tribes… most viewers won’t notice, but there’s a reason (e.g.) Lupita Nyong’o always wears green. This is great worldbuilding: it adds depth without getting in the way, and it rewards deeper viewing and re-viewing.

Edit: Gaze, if you dare, on Tom & Lorenzo’s overview of the costumes of Okoye, Nakia, and Shuri in particular. Ruth Carter deservedly got an Oscar for this.

Finally, a word on diversity, which is that it’s awesome. If you’re a Hollywood exec, rather than rebooting Batman for the 119th time, let some people tell stories that weren’t often given that chance before. The novelty and passion will make a better film. Also, trust me, give Shuri her own movie.




I finished King’s Bounty: The Dark Side. I already reviewed it, and pretty much everything I said then still applies, but I thought I’d record that I finally finished it, four years later.

One cute thing about the game is its occasional light jabs at other fantasy properties, e.g. this bit from an elven poetry contest:


Overall thoughts: if you’ve never played these, go play Armored Princess. But if you have, this one is just different enough to enjoy. I just replayed both. Which makes it sound trivial, but it’s not– they are sprawling games; DS alone takes about 60 hours.

Somewhat to my surprise, I found AP quite difficult this time– I could win the battles but kept running out of money to buy units. Fortunately, I learned that there are console commands! Hit shift-tilde; the key one is rage which refills your rage meter. You can type money n to get n gold pieces, and doublearmy n to double the units in slot n of your army. You don’t need these, but especially in DS, to keep refilling units means constantly teleporting around, and it gets tedious.

The heart of the game is its little battles… it’s hard not to talk about them without diminutives, because they’re cute and look like something in a board game. They are almost always winnable (unless you’ve chosen to fight enemies marked “deadly”), but the goal is really to lose as few of your own units as possible. This is harder in DS, since the units that can resurrect themselves and/or others are much rarer. But it’s fun to use lots of ranged units and high-damage tanks to decimate the enemy before they can even move.

The one thing I don’t like about the game (and the reason I’m not going to start another playthrough, at least for another few years) is the back-and-forth quests. These dudes always want something from another island, and not infrequently you have to interrogate everyone on that island to get it… or everyone in the game, unless you Google the quest.

The game also has some weird attitudes about women. (Note, it’s made by Russians, so they probably haven’t been challenged on this stuff so much.) On the one hand, its female units are badasses. The hardest fight in the game is against a few thousand Dryads– they can put your low-level characters to sleep, and they keep summoning new allies. One of the islands is Amazonia, which is suffering a rebellion of men, which you helpfully put down.  On the other hand, my character, the demoness, dresses like a dominatrix and gets a bevy of corrupted prisoners to wait on her. (But, well, your character is so small on the screen that you can almost always ignore this.)


I’ve had the impression that Overwatch League is getting more and more dull, one-sided games, and I decided to see if it can be quantified. It can.


I went through the current three stages, adding up the number of games with each possible score. Here’s the raw numbers.

Stage 3-2 2-1 3-1 3-0 4-0
1 17 7 29 2 15
2 13 4 25 2 26
3 7 2 19 3 16

I define “interesting games” as the ones in blue. Sometimes a blowout can be a really good game… but really, watching a far superior team wreck a worse one is not fun. A close score indicates a better match.

Note the progression: interesting games were 34% of all games in stage 1, 24% in stage 2, and 19% in stage 3 so far.

4-0 blowouts were 21% of games in stage 1, 37% in stage 2, and 34% in stage 3. Pretty striking trends in both categories!

Now, this may not be surprising. There were a lot of new teams in stage 1; players were getting used to their teams and learning about the opposition. But it seems it could have gone either way. There’s intense reason for mid-tier games to figure out what the top tier is doing. (Admittedly, the low-tier teams have it rough psychologically– a losing streak is hard to break.)

I expect there’s no solution to this, by the way. Maybe it’ll be the same next year, maybe not… it might take a year for other coaches to figure out how to break the Titans.

The overall standings are also interesting, as there’s such a sharp break into tiers.  The Titans, Excelsior, and Shock all have map W-L differentials of at least +45; the next best is the Gladiators with +17. Two teams (Mayhem and Justice) are doing absolutely awful; two more are under 1/3 W-L ratio. That leaves 14 teams in the muddy middle.

Edit: After writing this, I see the Titans just lost their first game of the year– to the LA Valiant, who are in 17th place overall. Now that’s an upset!

This stage, the Dallas Fuel has completely fallen apart: 1 W 5 L.  I feel bad for Jayne, my favorite streamer/commentator, who’s one of their coaches. Still, they made the playoffs last stage and are still at .500 overall.

It’s nice to see the Shanghai Dragons in 6th place this stage, with a winning overall record (10-8), since they lost every game last year.  But I’m pissed that they never play Geguri any more. Jeez, coaches, at least put her in during one map a game.


Time for another business report, dedicated to the patron saint of sales— St. François de Sales.

The occasion is that total sales for all books, over nine years, have just kicked over 30,000.  Here’s a breakdown by title:


As you can see, nearly half of sales (almost 14,000) are due to the LCK alone. Of course, it’s been out for the longest. A chart of just the last 3 months looks similar, except

  • the Syntax Construction Kit is a much bigger slice, on par with the other linguistics books and the PCK
  • the China and India slices are a tad wider; they sell about three times the rate of the novels

Print books make up 61% of sales, which is about the same as it was in the last report. Print is not dead!

What else can I tell you?

  • People either don’t care for hardcovers, or don’t know they can buy them. (You can get hardcover versions of the LCK and the Lexipedia. And it’s worth it; they stand up to constant use much better.)
  • About 15% of sales are from outside the US.
  • I’ve had 11 clients for whom I created conlangs.
  • The smallest slice in the above chart, the Historical Atlas, is 100 copies, which is pretty good for a fictional history.


Connie-5I received my second proof, and my name is spelled correctly on both the cover and the title page, so I’ve approved it! The printers are standing by, ready to roll press for you, the reader.

Here’s my description page, and here’s the Amazon page. If you have no idea what syntax is, or what a syntax book is, start there.

If you’re an e-reader, you’ll have to wait a few days. The Kindle version takes some extra preparation, as all the nice Illustrator diagrams must be converted to the high-tech formats of the early 1990s, GIFs or JPEGs. I’ll mention it here when it’s done. (Oh wait. Another page says they accept PNGs. That’s late ’90s!)

Don’t be a drag who has a two-page syntax page in your grammar. All the cool kats will be tripping on the real syntax in this book.


I just finished Language acquisition and conceptual development, edited by Melissa Bowerman and Stephen Levinson (2001), and I want to write down what I learned while it’s still fresh in my mind.

You may recall the book report on Everett & Jackendoff and their feud over innatism. The issue there is Chomsky’s longstanding contention that language learning is far too hard for children, therefore they must have a head start: grammar and vocabulary are already hard-wired into their brains. All they have to do is figure out which of a small series of switches to flip to get the grammar (“oh! verbs go last here!”) and work out that dog means Inbuilt concept #293291.

This book is a report from the trenches of language acquisition; if anyone knows how it goes, these people do. I note, by the way, that this is one of the few fields dominated by women: 20 of the 30 authors of these papers are female. Yay for linguistics!

There is no knockout punch— unsurprisingly, there’s a lot we don’t know about how children learn languages. And this book, at least, doesn’t have too much to say about how children learn syntax, much less whether they do so using Minimalism, Arc Pair Grammar, Role & Reference Grammar, etc. It’s mostly about the first three years, the first words learned, and what that tells us about children’s conceptual system.

The biggest news seems to be:

  • Children understand things far earlier than was once supposed. E.g. Piaget thought that children didn’t acquire the notion of object permanence till 3 years or so; we now know they have it at 5 months. He also thought that children didn’t understand the concept of time till about 8; but in fact they are clearly able to remember and refer to past events, and anticipate and refer to future events, at not much more than 1 year of age.
  • At the same time: universal, basic concepts are more elusive than ever. Languages really do divide up conceptual space differently, and this is evident in children’s speech from the beginning.

The object permanence result is due to better, cleverer technique: rather than relying on the baby’s actions, we only check what they’re looking at. Basically: babies can be surprised, and look longer at unexpected outcomes. So you show them a doll being placed behind a screen, then remove the screen. They’re surprised if they see no doll there, or two dolls.

Many of the authors refer to Quine’s problem. Quine envisioned a linguist eliciting words from a native. A rabbit goes by, and the native says gavagai. Does this mean “rabbit”, or “hop”, or “fluffy tail”, or “unspecified set of rabbit parts”?

Now, the linguists can’t bring themselves to say that Quine is just being a jerk. But there’s a pretty clear answer to this problem: we aren’t tabulae rasae; we’re animals with a hundred-million-year evolutionary history of perceiving objects, especially moving objects, and double especially animals. Some things are very salient for humans— we’re built to see rabbits as objects with a characteristic shape, size, and activity pattern. We’re not built to focus on rabbit tails or miscellaneous rabbit parts.

Early proposals were that children use some all-purpose generalizations: words are likely to refer to the most salient entities; words are normally not synonymous.

Going beyond this, there were assumptions that children would learn nouns before verbs, closed-class form words before content words, shape before materials, and that they would probably learn universal concepts first. This little list of assumptions turns out to be wrong: it depends on the language.

  • Many languages are far more verb-oriented than English. Kids still learn a lot of nouns, but sometimes the proportion of verbs is far higher.
  • Often very specific verbs are learned before abstract spatial words.
  • English children learn to pay the most attention to shape; Maya kids pay the most attention to material.

As for universal concepts, it’s worth looking in detail at an example provided by Levinson. The language is Tzeltal.

Pach-an-a bojch ta y-anil te karton-e.
bowl-put-cause.imp gourd at its-down cardboard-that

The intended translation is “Put the bowl behind the box.” But just about every detail in Tzeltal is different.

  • The shape and spatial information is largely encoded in the verb, not in nouns. Pach– means “place a bowl-shaped vessel upright on a surface.”
  • Corollary: the two NPs refer mostly to material. Bojch is really a word for a gourd; karton can refer to anything made of cardboard.
  • “Behind” is a relative term, which doesn’t exist in Tzeltal. Instead, an absolute frame of reference is used. “Downward” can refer to absolute height, but here it refers to horizontal location, because of a geographical particularity: Tzeltal territory is on a slope, so “downhill” also means “northward”.

Do children really master this system? Of course; they have a pretty good grasp of the slope system by age three. They also master a wide range of very specific verb forms rather than relying heavily, as English-speaking toddlers do, on “up/down”.

Another neat example: English toddlers quickly learn to distinguish “put ON” from “put IN”. Korean children divide up this semantic space quite differently, using at least seven verbs.

  • kkita means “fasten tightly”– this includes putting the top on a pen, placing Lego bricks together, putting a piece in a puzzle, placing a cassette in its box, or buttoning a button.
  • nehta means “place loosely”– e.g. put a book in a bag, or a toy in a box.
  • pwuchita is used for juxtaposing surfaces– e.g. placing a magnet on the fridge.
  • nohta is used for placing things on a horizontal surface.
  • for clothes, you have ssuta for hats, ipta for the body, sinta for the feet.

All this is fascinating because philosophers and linguists are apt to take English categories and assume they are universal concepts: UP, DOWN, IN, ON. Nope, they’re just projecting English words onto Mentalese. There is no stage where children use “universal” concepts before using language-specific ones. (Indeed, there’s evidence that children understand the language-specific concepts well before they can say the words.)

Does all this “affect how you think”? Of course. Levinson tells an amusing anecdote: he almost got his truck stuck in quicksand when his Australian Aborigine companion told him to “swerve north quick”. Levinson just couldn’t calculate where north was fast enough.

There’s also interesting tidbits like, did you know that there is a gradient between comitative and instrumental? It goes like this:

1 – give a show with a clown
2 – build a machine with an assistant
3 – captured the hill with his squad
4 – performed an act with an elephant
5 – the blind man crossed the street with his dog
6 – the officer caught the smuggler with a police dog
7 – won the appeal with a highly paid lawyer
8 – found the solution with a computer
9 – hunted deer with a rifle
10 – broke the window with a stone

In English, as you can see, we use “with” for all of these. In a multitude of languages, these meanings are divided up linearly. E.g.

  • Iraqi Arabic: 1-8 vs 9-10
  • Swahili: 1-6 vs 7-10
  • Slovak: 1-9 vs 10
  • Tamil: 1-2 vs 3-10

That’s pretty neat!

Anyway: there’s still a lot of argument on how exactly children learn, whether they start with particular cognitive abilities, whether they have particular linguistic abilities. Many authors point out that innatism doesn’t really help reduce the problem. E.g. to see if dog matches Inbuilt concept #293291, you pretty much have to have a sense of what a dog is. If you have that, what good is the inbuilt concept?

You could try to save innatism by multiplying the number of inbuilt concepts. E.g. you include the 10 steps of the comitative/instrumental gradient, and both Korean and English positioning concepts, and both English and Tzeltal directional systems. But this is only complicating the child’s problem. Rather than finding quick matches between the words they hear and a small number of universal concepts, they have to consider hundreds or thousands of alternative conceptual systems.

It’s also worth pointing out that parents are far more helpful than Quine’s native informant. People don’t just say words at random. As Michael Tomasello emphasizes, language is often presented as a commentary on a situation the child already understands, such as moving toys around with her mother. There’s a lot of repetition; the parents’ language is emphatic and simplified; the parents are not trying to confuse the child with talk of bags of rabbit parts.

BTW, this is in theory the last book I’m consulting for my syntax book.  So, I’ll soon have a first draft, at least.


Next Page »