Comic strips have developed an iconography for cursing. Here are a few samples:

Mort Walker, the originator of Beetle Bailey, calls these icons jarns, quimps, nittles, and grawlixes. Linguist Arnold Zwicky prefers to call them obscenicons. In the late 1950s, looking at our local paper, the Cortland Standard, I noticed that standard spelled backwards is dradnats. I thought that dradnats would be a good name for the cuss-word surrogates. When I was working on my comics manuscript, I collected many examples from comics old and new, and posted them. They go back as far as 1909, if not farther.

I found that French grawlixes are more elaborate, on the average, than the American ones. They contain symbols of violence such as fists, skulls, and weapons more often. The closest French equivalent for “seeing stars” is “voir trente-six chandelles” (literally, to see 36 candles). This may account for the prevalence of candles in the iconography of French grawlixes. Pseudo-Chinese characters are also common there.

To approximate grawlixes on a typewriter keyboard, mostly use the shifted number keys, e.g., !@#$%^&*. It is convenient for writers of prose that cartoonists came up with an iconography that’s so easily adaptable to typing.

The linguists who post on Language Log tend to deprecate taboo avoidance, equating it with a belief in word magic. I think there should be a small set of words that show that the speaker is so upset that he or she is willing to break taboos over it. Words are better than fists for expressing strong feelings. They have to be used very infrequently. Some taboo words are used so often in movies that their power to shock is greatly diminished.

I personally dislike hearing offensive words. I try to avoid offending my audience, too. There is no contradiction here. Only offend intentionally, and it is my not intention to offend.


Another Harebrained Idea

When I wrote this song (1970), I was renting a little cottage out in the woods. It seemed like a cute idea at the time, but, frankly, I never put it into action. Since then, I think I’ve heard other music with about the same tune, so maybe I borrowed it subconsciously.

Let us have a little power failure,
Power failure of our very own.
Let us have a temporary blackout.
It will make us feel so all alone.
Turn the lights out; I'll turn off the furnace.
Off, the radio and TV too.
Then we'll have our homemade power failure,
Only big enough for me and you.
I'll pick firewood, you can hold the flashlight.
Then as we sit by a cheery blaze,
I might put my arm around your shoulder.
For your back's warmth you would let it stay.
Summer is the time of play and laughter,
When the day and future are so bright,
But for us, the future can be brighter,
On this crystalline December night,
On this crystalline December night.


A Benefit from Difficult English Spelling

As is well known, Spanish spelling is much more phonetic than English spelling. This fact has an unexpected consequence. It is also well known that people who sound out their words and subvocalize them as they read are much slower readers. English speakers are obliged to do some form of look-say reading. They can’t sound the words letter by letter as a rule, because of the many inconsistencies of English spelling. Spanish speakers take the short cut of reading by sounding out the words. As a result, they tend to read slower.

In my computational linguist work, I had the opportunity to tabulate the kinds of spelling error most often made by students. In Spanish, the errors were overwhelmingly caused by phonetic spelling. The letters ‘b’ and ‘v’ are pronounced the same in almost every dialect of Spanish. So are the letters ‘c’ and ‘z’, and so are ‘ll’ and ‘y’. (If I wanted to make those statements more precise and general, it would take up a lot of space, so please accept them as approximations.) So the kind of error that you expect to see a Spanish child make would be writing bibir for vivir, or Me yamo for Me llamo. Foreigners are less likely to make those mistakes.

Maybe the take-away is that if you have to work extra hard to learn something, it can pay off in unanticipated ways.

I have disavowed political correctness, but nothing in this post should be taken to imply that Spanish speakers are dumb. They might read English as fast as anyone.


What a Computational Linguist Does

For eight years, I worked as a computational linguist. When people asked what I did, they often wanted to know what that meant.

Academic linguists come to expect that people who meet a linguist will ask, “How many languages do you speak?” Being a linguist is not about speaking many languages. It’s about discovering the implicit rules that speakers of a language follow unawares. (The “about me” page answers that question, though.)

Computational linguistics applies “big data” methods to do natural language processing (NLP). NLP applications include machine translation, spell-check suggestions, autofill suggestions, search engine page ranking, sentiment analysis on social media, recognition of voice input, question answering, essay scoring. As you are probably aware, all of these functions are still far from perfected, but IBM’s Watson program did pretty well on Jeopardy!.

A rough description of “big data” methods would go like this. Start with thousands of complex entities. Use human intelligence to categorize a limited subset of them. Select some of their properties that can be measured by a computer. Categorize the rest of the original entities by finding what categories their nearest neighbors belong to, where “nearest” means the ones whose measured properties as a group are closest.

As a rough analogy, say that the entities are cities of the U.S., the categories are states, and the measured properties are latitude, longitude and altitude. If you know the measurements of an unknown city, and all of the cities that have nearly the same measurements are in North Carolina, the unknown city is probably also in North Carolina.

The entities a computational linguist deals with are usually text samples, but they can also be recorded spoken words. In my job, I dealt with text. One type of contribution I made was to think of new measurable properties, and then to do the programming to measure them. Most of the measurables involved looking up the words of the text in an electronic dictionary, which had attributes such as number of syllables, part of speech, and so on for each word.

I would sometimes say that to get my job, I had to know more than a little about mathematics, statistics, programming, languages, and grammar, but not as much as an expert in any of those fields.


Latin Schmatin!

I was chatting with a fellow choir member about singing in various languages, and how to pronounce them.

There are at least three established systems for pronouncing Latin. You just have to pick one.

When I was taking Latin in high school, we learned what is often called the classical system. The letter ‘v’ was always pronounced ‘w’, ‘j’ was pronounced ‘y’, and ‘c’ and ‘g’ were always pronounced hard. Thus, ‘veni vidi vici‘ came out “way-nee, wee-dee, wee-kee,” and “Julius Caesar” was “Yoo-lee-us Kye-sar.”

The classical system was developed by philologists, mostly German, attempting to reconstruct the way the Roman senator Cato would have spoken in the second century B.C. There are no recordings of Cato’s speeches, and word-of-mouth transmission is not reliable over a span of twenty centuries. In fact, word-of-mouth is what produced modern Italian. How did the scholars find out how Cato spoke? There is evidence in the rhymes found in ancient poetry; more evidence in the way Greek words were respelled when quoted in Roman text; and still more in what happened to Latin words when various national groups adapted them to their vernaculars. There is more direct evidence in ancient language instruction manuals.

Around the end of the 19th century, British schools taught a Latin pronunciation as if the words were English. Around 1870, pedagogues at Oxford and Cambridge urged the teaching of the classical system. Like most teaching reforms, this was controversial. James Hilton used the controversy as a plot element in his 1934 novel “Goodbye, Mr. Chips.” For example:

“… This question of Latin pronunciation, for instance–I think I told you years ago that I wanted the new style used throughout the School. The other masters obeyed me; you prefer to stick to your old methods, and the result is simply chaos and inefficiency.”

At last Chips had something tangible that he could tackle. “Oh, that!” he answered scornfully. “Well, I–umph–I admit that I don’t agree with the new pronunciation. I never did. Umph–a lot of nonsense, in my opinion. Making boys say ‘Kickero’ at school when–umph–for the rest of their lives they’ll say ‘Cicero’–if they ever–umph–say it at all. And instead of ‘vicissim‘–God bless my soul–you’d make them say, ‘We kiss’im’! Umph–umph!”

Another system, which I call the church system, is to pronounce Latin words as if they were modern Italian. The hierarchy of the Roman Catholic Church has tended to be composed of Italians for centuries, and their pronunciation has been influenced by their native language.

The first time I was in a choir singing “propter magnam gloriam tuam,” we were instructed to sing “mahn-yahm.” I was shocked, because I had been taught the classical system and had learned “mahg-nahm.” But it made sense, because the piece had been written for a Catholic mass.

That leads us to another system, sometimes called the Continental system, in which Latin words are pronounced as if they were German. When we sing a piece in Latin by Bach, Beethoven, or any German composer, we usually use that system. It’s also an appropriate system for pronouncing Latin names of species in biological taxonomy.

In this matter, as in many others, people become attached to the methods they learned with great effort and have adhered to for years of their life, often without realizing that there are alternative methods that are equally valid.

It’s not an issue for choirs and scientists only. We use Latin-derived words and phrases in everyday speech, such as alumnus/a/i/ae, bona fide, prima facie. Advice to lawyers: if you’re in court and the judge mispronounces a Latin phrase, be judicious. Tactfully mispronounce it the same way. (Note: bona fide, ablative case, means “in good faith,” and bona fides just means “good faith,” not a plural.)


The Names Not Used

When Janice and I had three sons in the Law family and were expecting a fourth child, we were on the lookout for a good boy name. We had already chosen some good strong ones for the first three. I suggested Walter Everet Law (sic), because it’s a palindrome. We would instruct our first three children never to mention the special nature of the name. Someday, Walter would come home from school, all excited, and say, “Guess what? My name is the same forwards and backwards!” We would all laugh together.

If we had had triplets, I was prepared. Wallace Cal Law and Waldorf Rod Law are also palindromes. But we are grateful that it was a girl instead. We don’t really like the nickname Wally.


Publish SSNs

You know that you have to protect your Social Security Number to avoid identity theft. But there are many places where you’re legally required to provide your SSN, so the government can keep track of movements of money. I stopped carrying my Social Security card years ago on advice from experts, but now I have to carry my Medicare card with me, and it has my SSN clearly visible.

There are many organizations that want to have a unique identifier for everyone they deal with. It’s even better for them when all of those people know their own identifier by heart. Your local hospital might want to give you a seven-digit case number when they admit you at the ER, but they know that you have enough numbers to remember already, and they can’t expect you to memorize another one for their convenience. Your name won’t do, because there are too many people with the same name. Nowadays they ask for date of birth, too, because they’ve learned from bitter experience that the name is not enough.

What is the problem with using your SSN freely as your identifying number everywhere? Apparently the problem is that some financial institutions accept the SSN as proof of identity. That’s insane. It’s kind of like saying that as long as you know a car’s VIN number, that proves it’s your car.

The government should forbid accepting the SSN as a proof of identity. It wouldn’t even take an act of Congress. Just publish everyone’s name and SSN. That would make it obvious to even the most obtuse institution that they can’t use the SSN as positive identification. Don’t give my money away to anyone just because they can tell you my SSN.