A constant reader pointed out this video to Dr. Boli, thinking he might be amused by it. It is a list of things retirees should do to protect their homes, and it is probably all very good advice. Dr. Boli heard none of it, because he was much more interested in the AI-generated art.

The art all had a consistent style—homey neorealism with crusty brushstrokes meant to suggest that it had been done by a good amateur painter. It worked well for the subject—except that many of the images were meant to represent documents or other things with words on them, and that was where the art became strangely and unintentionally surreal. Bot artists still have a lot of trouble with words.

That seems strange, because AI has become very good at optical character recognition. It is the one thing Dr. Boli regularly uses AI to accomplish: the bots can read a page written on a hundred-year-old typewriter and transcribe it without a single mistake.

But when the traffic is going in the other direction—when the bot has to make words, not by generating Unicode characters, but by actually drawing pictures of letters—then the results are comically inept.

It appears that our videographer has edited some of these images to make the primary phrase correct, but otherwise he has let the AI bot give us its impression of what English looks like. And its impression is very impressionistic. It is not just that the letters spell out words that do not exist; if you look closely, you begin to see that some of the letters are not real letters in any known language.

This is what human written language looks like to an alien intelligence—for if we stipulate that the bots are intelligent (which not everyone will agree to), then we must at least call them alien intelligences.

Sometimes the bot has obviously been given directions about the words to include in the image; but the bot has felt free to treat the words as impressionistically as the pictorial part of the image.

Prompting, it seems, can generate very real-looking pictures, but these attempts at written language probably tell us something very interesting about how much the bots actually understand of what they are picturing for us.

Even this SOLD sign, which gets the word right, includes what is obviously meant to be a phone number that looks like no phone number in use today—although, in an alternate universe where named telephone exchanges were kept into the YouTube age, this number would make sense as “STanton 42-4507.”

So what have we learned from our little excursion into botdom? First, that unsupervised bots need to be reined in and their work refined or altered by humans if the humans do not wish to be embarrassed. Second, that there seems to be no penalty for extreme laziness in video making, so go ahead and upload your first attempt.

And What We Can Do About It.

An American soldier making himself understood in Romanian.

---

Why don’t most Americans learn a foreign language?

Cynics will tell you that it is because Americans are lazy ignorant xenophobes. And the cynics are mostly right, if only because most human beings are lazy ignorant xenophobes, and Americans are no better than anybody else.

But many lazy ignorant xenophobes in other countries are bilingual or trilingual. Some of the most profoundly stupid people Dr. Boli has ever known have been fluently stupid in two or three languages.

So why not Americans?

There are practical reasons. We live in the midst of an ocean of English, and few Americans need to understand any other language to get on with their everyday lives. It is true that they may have thousands of neighbors who speak Spanish or Korean or Uzbek; but average English-speaking Americans deal with those neighbors by ignoring them most of the time, and by shouting at them when they cannot be ignored, on the well-known principle that shouting makes even foreigners understand plain English.

But Dr. Boli proposes another reason as well, which is that our methods of teaching languages make learning them impossible for all but the most motivated students. The pedants have taken over language instruction, as they have everything else, and they place prickly severe-tire-damage roadblocks right at the beginning of the course to discourage us from learning anything.

The very first thing you have to learn in a language is the set of basic sounds that language uses. Native speakers of a different language will probably use sounds you cannot imitate. The pedants will insist at the beginning of the course that you must spend a week trying (and failing) to get those sounds right. After that week you are thoroughly discouraged; you say “This is hard” and abandon the class and sign up for Macrame 101 instead.

Consider the French nasal vowels. You want to ask where the restaurant is in French: “Où est le restaurant?” It’s as simple a question as you could ask, but that word “restaurant” ends in a sound that American mouths cannot form—partly because it involves not just the mouth but the nose as well. You are stuck at the beginning of your Intro to French course, and you feel like a fool.

But now think back to your acquaintances who are fluent in multiple languages. Dr. Boli has had good friends from Spain, Hungary, China, Germany, Ethiopia, and many other places who all spoke English fluently: they could express any idea easily, with grammatical precision and expressive elegance that would shame most American native speakers of English.

But they spoke with an accent. You could tell the German was German; his vowels and consonants came out as approximations of English sounds adapted to the way a native German could easily form them. And that was fine. No one had any trouble understanding him. We noticed his accent at first, but we forgot it a few minutes into a conversation with him. The same is true of all the others: they made themselves understood perfectly in English, perhaps even better than many native Americans.

Successfully bilingual people are usually people who have got over their fear of having a funny accent.

So we go back to our American introduced to French for the first time. He wants to ask where the restaurant is. We can give him diagrams of mouth positions; we can show a cutaway view of the human head with arrows indicating the direction of air flow. He will be confronted with a pneumatico-anatomical treatise that will make him despair.

Or we can just tell him to say this:

“oo AY luh RESS-to-RAHNG?”

Yes, he will have an American accent in French. But any French speaker will understand that question and be able to point to Chez Micheline on the corner.

In the Second World War, American soldiers descended on Europe and Asia by the shipload, and they had to learn in a hurry how to make themselves understood. This was an urgent necessity, and fortunately the War Department had the best technical writers ever gathered together in one enterprise. They put together quick guides for the soldiers that would let them make themselves understood, funny American accent and all. KAWM-pruh-nay VOO?

These are probably the best introductions to foreign languages ever written for American English-speakers, because they assume that the goal is not native pronunciation but communication. The pronunciations shown in the guides are not perfect: they are simply good enough.

You will find all the words and phrases written both in French spelling and in a simplified spelling which you read like English. Don’t use the French spelling, the one given in parentheses, unless you have studied French before. Read the simplified spelling as though it were English. When you see the French word for “where” spelled oo, give the oo the sound it has in the English words too, boot, etc. and not the sound it has in German or any other language you may happen to know.

These manuals were meant to accompany a set of records, and if you were the lucky kind of soldier who could sit down with a phonograph and listen to the recordings, you might hear the actual native pronunciation of the language. But the compilers of the manuals knew that many soldiers would not be so lucky, so the printed pronunciations would have to do the job well enough.

Longtime readers may already have guessed this: Dr. Boli has rounded up most of these War Department language manuals and created a new page in the Eclectic Library:

War Department Language Guides.

A while ago we were talking about right-justified text, and in particular about the clashing dogmas on the subject: Web designers on the whole think right-justified text is a mistake in any context, whereas designers of serious books and magazines think that failing to right-justify text is a mistake. We noted that WordPress, the software that has run this Magazine since it moved to the Web almost nineteen years ago, and that since that time has grown to power half the Internet, had removed the right-justification feature years ago, leaving designers to write their own CSS if they felt the need for a traditional text design. (This was one of the main reasons Dr. Boli had to learn a little about CSS and theme design in the first place.)

Yesterday we were reading about the new features of the upcoming version of WordPress, and buried two-thirds of the way down a very long page was this little paragraph:

Nearly all text blocks now support the standardized text-align block support system, including Paragraph, Button, Comment blocks, Heading, and Verse. Plus, text justify alignment is now available. See tracking issue to follow along on the progress (60763).

And, indeed, it seems that a pull request that has been open for two years has finally been closed with a successful merging of the code into the core of WordPress. It is now possible to justify text with a simple button, even if the theme (unlike the “Dr. Boli” theme that runs this Magazine) has not specified justified text.

Throughout history, civilization has been under constant assault from the forces of barbarism. But once in a while it helps to remember that, in spite of tremendous defeats, there have always been more victories for the forces of civilization, and the barbarians have never succeeded in completely extinguishing it. Today we are witnesses to a victory, and we can remind ourselves that defeats can be overcome and pernicious trends reversed if someone with enough determination and patience stands up for civilization.

If justified text is finally back, could small capitals be next? The “Dr. Boli” theme has made them as easy as possible under current circumstances, but they still have to be implemented by adding the codes <sc></sc> directly in the HTML. It seems to Dr. Boli that simple and intuitive small capitals, which are required for certain uses in formal publishing, are the next obvious step in making the World-Wide Web safe for serious writing. Is anyone ready with a pull request? For any coders who need encouragement, it may help to point out that the successful pull request for a right-justification option came from a first-time contributor to WordPress.

Cannellini (noun, from the Italian).—Small Italian cans, commonly used to package prepared beans for the retail market.

Transcribed below…