How Many Parts of Speech Are There?

In school, we learn that there are eight parts of speech: noun, verb, adjective, adverb, pronoun, preposition, conjunction, and interjection. Modern grammarians recognise a couple of others: determiners and particles.
Phrasal verbs are idioms combining a verb and a particle. Some examples are “slow down,” “stand up,” “turn over.” Those particles (down, up, over) all look like prepositions, but that’s not how they function.
In “Jack and Jill went up the hill,” “hill” is the object of “up”. In “Stand up when a lady enters the room,” there is no object of “up.”

Determiner is a useful basket category that includes articles, and many types of adjectives, such as demonstratives (this, that, these, those), and quantifiers (some, many, all). The reason for making them a separate category is to simplify the statement of some grammar rules. In my computational linguist job, I found that I could get more precise and significant parses of sentences by using those parts of speech. The determiners that you can use with mass nouns (such as milk, patience, and time, are not always the same ones you can use with count nouns (such as book, event, and second). For us, part of speech was just a convenient arbitrary label, to be used in whatever way improved the analysis.


Hey, Google! Pick me!

In one of my several periods of unemployment, I dabbled in SEO (Search Engine Optimization). One of my neighbors had an e-tail website. He asked if I would help him raise his site’s prominence on Google. I read everything I could find
on the subject. One rule is to provide the best content. If you build it, they will come. The search engines have developed very sensitive ways to measure the respectability and authority of a site, and how well it satisfies their users, based on the search terms. It’s like a courtship between the search engine optimizers and the search engines. The website tries to put on its most attractive face. The search engine doesn’t want to seem “easy.” A search engine optimizer has to be aware of the tricks that other SEOs are using, and that Google has learned about and taken countermeasures against.

At that time, my own site,, had the very respectable PageRank of 5, sometimes straying as high as 6. The PageRank is based on how many other pages link to yours, and how respectable those pages are. So one effective SEO strategy is to persuade lots of other webmasters to link to your site.

You can pay a search engine to boost your site in the returns. I rejected that approach. I felt that offering superior content was more honorable and probably more sustainable. A pitfall to avoid for this site was that he was selling a skin care product. If he made any health claims, the FDA was poised to come down on him like a ton of bricks.

Company Theme Song

About 1990, the CSX railroad and American Airlines found that they were handling multi-modal shipments together. They each had their individual electronic tracking and tracing systems. It was a natural move to link those systems together. To do so, they formed a joint venture called Global Logistics Venture (GLV). Over the years, it had its ups and downs, and changed its name to Encompass and later BridgePoint. We linked to other carriers, including ocean and trucking.

I was one of the first employees, doing programming and analysis. I was part of a skunk works project. We put together a large database showing all the movements of goods that any of the carriers in the project was able to perform, and used it to generate an itinerary for any shipment.

For the 1994 annual company meeting, my boss, Rick Poff, wanted to put on a dog and pony show.He asked me to write a theme song for it. I said sure, if I could do it on company time.Rick loved it, and asked me for the sheet music again and again. For the  show, he cobbled me a container costume. It was a big box hanging from my shoulders on suspenders.About a dozen of us sang this song:



  1. I’ve got a load of freight to move across the state.If it’s not there tomorrow, There’s  a fine. I’ve got to move this crate, and it can’t get there late, or else my job will lay right on the line. Encompass, tell me where my cargo’s gone, Is it en route or is it on my lawn? Please say it’s almost there, ’cause if it’s not, I’m in a pot of water and it’s getting hot.
  2. . I’ve got a full truck load that should be on the road. It’s due in North Dakota in a week. I sent a brand-new chair to Singapore by air, And when the chair got there it was antique. Encompass, let me see your crystal ball. Let me say, “Mirror, mirror on my wall.” Encompass, let me rub your magic lamp. I’m not a chump if you can help me be a champ.
  3. I’ve got a full container bound for Transylvania on the Lusitania, and I hope She doesn’t hit a reef, she doesn’t come to grief; it’s full of iron bars, not Ivory Soap. Encompass, help me pick which goods to ship, and find the route that’s fastest for the trip.Each carrier summon to a rendezvous, and track and trace and see how good a job they do. If you can do all these, Logistics is a breeze. I want all this and more in ninety-four!
    For the 1995 company meeting, we reprised the song, updating the last line, to go “We’ve got the basics down, it’s time to go to town–the future will arrive in ninety-five!”

The Failure of “Candide”

Voltaire’s Candide is one of the most celebrated philosophical novels of all time. It is generally understood as an attempted refutation of Leibnitz’s optimism. Leibnitz’s position is usually summarized, “All is for the best in this best of all possible worlds.”
We all know there’s a lot wrong with the world, so what Leibnitz is saying can only mean that if you adjusted something to fix one problem, that very adjustment would cause other problems that might be even worse. In other words, that alternate world is not even as good as this world.

Any debate on this question can only be speculative. We can’t perform experiments. We don’t have the power to make marginal changes to the world that fix specific problems. If we did, and new problems popped up, how would we know whether they were consequences of the changes we made?

Voltaire’s response is not much more than a flat-out contradiction, without supporting evidence. He shows us some truly awful things happening in this world. He expects us to conclude that there must be a possibility of a better world somehow. He just doesn’t give us any details on how that could come about. A world without devastating earthquakes would be an improvement. Is that really possible? (Spike Jones and his City Slickers did a cover of Ghost Riders in the Sky, where they sang “On horses snorting fire!” Slicker: “Is that possible?” Jones: “How would I know?”) Same here. Earthquakes seem to be a consequence of plate tectonics. Is it possible to have a planet teeming with diverse forms of life with a solid unitary crust over a long time span? How would Voltaire or Leibnitz know?
The progressive   mindset is influenced by Darwinism and is based on the assumption that things get better and better over time. That mistakes don’t last long. They die out by natural selection. Candide doesn’t even make that argument. Voltaire seems to make it an article of faith. And I don’t see how natural selection could cause earthquakes to die out.


A Lecture by H.S.M. Coxeter

The eminent geometer H.S.M. Coxeter (not to be confused with a Gilbert and Sullivan operetta) gave a colloquium at Caltech on Feb. 8, 1977. I attended it, and while it was fresh in my mind, I wrote the following account. Coxeter came into the lecture hall amid a bunch of students. They had probably just had a departmental tea. The students were grungier than I expected. At Johns Hopkins, male students were expected to wear a tie to colloquia.

Coxeter looked at the usual empty seats in the front rows, and invited anyone to come closer and get a better view of the blackboard. He was bald on top, fringed with cute curly white hair, not large in stature, and continually grinned in an engaging way. As he paused, no one came forward. He repeated the invitation. One student got up and moved to the front, and the rest all applauded.

He spoke about certain patterns of numbers and their geometric relationships. It was all very easy to follow; he’s a good teacher, as I heard some of the students comment after the lecture. He did absent-mindedly make a few mistakes. He caught some of them. I never correct anyone else’s slip of the chalk, because it doesn’t impede my own understanding, and if it did impede another listener’s, it was up to that other to ask. Nit-picking just slows down the lecture.

At one point he had us all laughing. There is a number that crops up in innumerable ways. It’s φ=(1 +5)/2, about 1.618. It’s called the golden section, mainly because if you take a rectangle whose sides are φ and 1, and cut off a 1×1 square, the remaining piece is the same shape as the original rectangle. That is, (φ-1):1 :: 1:φ.


Φ appears again in the Fibonacci series-the numbers 1, 1, 2, 3, 5, 8, 13, …, where each number is the sum of the previous two. The Fibonacci series developed from the question, if one pair of rabbits produces another pair of (baby) rabbits every month, and baby rabbits are ready to breed in their turn after two months have gone by, how many pairs of rabbits will there be after so many months? But Fibonacci numbers crop up in all sorts of unexpected ways. The number of seeds in one spiral of a sunflower or the number of spines around a pineapple is almost always a Fibonacci number, for instance. Now, it turns out that the ratio between consecutive Fibonacci (“fee-bone-otchy”) numbers approaches φ. The first few ratios are 1/1, 2/1, 3/2, 5/3, 8/5, 13/8, …, which as decimal numbers are 1, 2, 1.5, 1.66666…, 1.6, 1.625, …; the farther on you go, the closer you get to 1.618.

In his talk, Coxeter alluded to the golden section, which is the length of a diagonal of a pentagon whose side is 1. He called it τ (the Greek lower-case tau), even though most people call it φ (the Greek letter phi). I pronounce φ “fie,” but many equally erudite people pronounce it “fee”). The Greek letters ξπφχψ are xi, pi, phi, chi, and psi. I aim for consistency, and rhyme them all with “pie” for π, and I don’t hear anyone calling π “pee.” Coxeter explained that τ is for τομοσ, which is Greek for ‘cut.’ He joked, “People who call it φ are just making a feeble pun on Fibonacci,” and everyone laughed. I remember one other risible moment in a math colloquium. One problem in abstract algebra is the classification and enumeration of groups. The visiting professor strove mightily and concluded, “So there are exactly eleven groups of this type.” The students laughed, because they realized that it had been ages since anyone had mentioned a number lower than twenty in their classes or lectures. As a rule, math grad students don’t deal with specific numbers; they represent them with letters or other symbols. Besides, most classes of groups that we knew had infinite numbers of members: the cyclic groups of order n, for any positive integer n; similarly for the symmetric groups, or groups of permutations on n elements, again for any n.