The Keyest Concept
Posted by Neal on November 4, 2009
“The discussion we had yesterday,” I began, “was a reminder to me that language data is always messy. I was trying to show you a simple picture of how parts of speech worked, and you guys kept giving me words that messed up the nice picture I was trying to paint for you. Language was invented over thousands of years by millions of people, so there are going to be exceptions, and words that you can’t easily label as one part of speech. That’s just the way it is. The good news, though, is that the tests we’re doing here are tests that you can do on your own, so you can see how a particular word is behaving.”
“Yesterday was also a reminder to me,” I continued, “that you can’t rely on just one test to determine what family some word belongs to. For example, yesterday someone gave me the word cheesy to replace the in the lazy dogs. But you probably have a feeling that cheesy isn’t like the other words we called determiners. The, a, every, some, cheesy — one of these words isn’t like the others. How can we prove it’s not like the others?”
From there I pointed out that cheesy could not replace our first the for *cheesy brown fox. Well, that didn’t prove anything. The same could be said for these and those, but we still called them determiners.
However, look at what you could do with cheesy: You could say cheesier and cheesiest, and This is really cheesy! “Some of you probably have said that,” I observed. You couldn’t do any of that with words like the or every. *The-er and *The-est were just nonsense, as was *This is really the. So although cheesy was able to replace the in our first sentence, all things considered it was behaving more like an adjective.
I wasn’t done with adjectives yet. One student the day before had asked: Wasn’t any noun that described another noun really an adjective? That was another issue that deserved some attention. I pointed to a spot on the whiteboard. “Over here Mrs. _____ has written two objectives: Knowing how to use parts of speech, and knowing what a complete sentence is. She said they were key concepts, and I wrote that down right here next to them. Not because they’re key concepts — although they are — but because this is an interesting term. How many of you would say key here is a noun? Like, these concepts are like keys that will give you access to important knowledge?”
A few students raised their hands.
“How many would say key is an adjective that means ‘important’?”
A few other hands.
“How many just don’t know?” Plenty of hands now.
“A lot of people just don’t know. Either an adjective or a noun could fit there. If you decide key is an adjective, what might you be able to do with it?” We put key through its paces: keyer/more key, keyest/most key, and This is (really) key. Just about everyone hated the comparative and superlative forms of key, while most were OK with key as a predicative adjective (…is key). So key behaved in some ways like an adjective, but not all of them. Messy, yes, but that’s just the way it is. You can have it simple, with a nice short list of eight (or however many) parts of speech, and then wonder why some word was lumped in one class when it didn’t quite fit there. Or you can have a more complex picture, but with the comforting knowledge that you can investigate for yourself how a word behaves, and see which other words it behaves the most like.
With that repair work done, I moved on to the idea of certain strings of words sticking together more tightly than others, and we went through carefully selected strings of words, applying what syntacticians know as constituency tests. I started by asking for single words that could replace the quick brown fox, and got he, she, it, and a proper name. Then for the lazy dogs, we got them. Moving on to over the lazy dogs, the students gave me a variety of prepositions that could act alone as adverbs: away, down, up, etc. They weren’t what I had in mind, but I guess they still illustrated the point. What I had in mind was there (or thither if you want something specifically with a directional meaning component and don’t mind an archaism, or thataway if you consider it to be one word). Then we expanded to jumped over the lazy dogs. The word I had in mind was did, and I was able to elicit it by asking, “Who jumped over the lazy dogs?” and asking for a slightly longer answer when they said simply “The quick brown fox.”
Moving back to over the lazy dogs, I introduced the movement test: Could the cluster be placed somewhere else in the utterance? I suggested Over the lazy dogs the quick brown fox jumped and Over the lazy dogs jumped the quick brown fox. One student said that sounded like something Yoda would say, so we took a digression into Yoda-speak. The student offered Defeat Darth Vader you must, which was a nice piece of evidence that even within must defeat Darth Vader, we had the smaller, moveable cluster defeat Darth Vader. Plus, I got to impress Doug’s classmates with my Yoda voice.
The last clusters I wanted to identify were quick brown fox and lazy dogs, and for them I brought in the wh-replacement test. “If I said to you, ‘Hey, guess what! The [mumble mumble] jumped over the lazy dogs!’, what would you say to me?” After a few false starts (“I can’t hear you” and “What?”), I managed to elicit The WHAT jumped over the lazy dogs? (and shortly later, The quick brown fox jumped over the WHAT?).
With all those clusters carefully identified, I asked for strings of words that didn’t pass our tests. It was tougher than I’d figured on. They discovered you could replace The quick brown with just a determiner, and with a wh-word: Which fox jumped…? I couldn’t remember subtler tests that would put quick brown into a cluster with fox, and I didn’t have time for them anyway. I just said that maybe the quick brown was a cluster after all. For jumped over, the students realized you could put in any transitive verb. On this one, I said that there probably were speakers for whom jumped over really did form a tighter cluster than over the lazy dogs, and that the same kind of clustering probably contributed to sentences like This bed was slept in by George Washington, Elvis Presley, and Michael Jackson. “I want that bed,” one girl said. “It’s not for sale,” I told her.
Then we tried out The quick brown fox jumped. One student correctly observed that the sentence I’d used earlier, Over the lazy dogs the quick brown fox jumped, could serve as a movement test for the quick brown fox jumped as easily as it could for over the lazy dogs. I didn’t have time to probe this cluster any further, so I acknowledged the point, but stuck with the idea of over the lazy dogs as the cluster, since it also passed the one-word replacement test.
Finally, I went through the sentence and put tents and labels over all our clusters: noun phrase, prepositional phrase, verb phrase, and (for quick brown fox and lazy dogs) nominal. When we got to the verb phrase tent, I highlighted how some verbs could make a VP all by themselves, like slept or jumped, while others had to have other stuff with them to make a VP, like tried or believed. But they were still all called verbs because they all formed VPs, which behaved similarly as a family. The last cluster we labeled was the entire sentence, consisting of the noun phrase the quick brown fox and the verb phrase jumped over the lazy dogs. (I’m hoping that the kids will remember this kind of sensible and informative kind of phrase diagram when and if they ever have to do Reed-Kellogg type diagrams, and know that there’s something better out there.) And now for the payoff for the whole two lectures: What did parts of speech have to do with complete sentences? Parts of speech combined to make phrases; phrases combined to make bigger phrases; and a sentence was one particular kind of phrase. The kind we’d made was a noun phrase plus a verb phrase — a rule for making sentences that they’d probably learned with different names: Sentence = Subject + Predicate.
“That’s not the only rule for making a sentence!” I hastened to add, but it was the most common one in English. For an encore, I talked about because. “Just like over took this noun phrase and turned it into a prepositional phrase,” I said, “because takes an entire sentence and turns it into something else, a phrase that acts like an adverb. It’s not a sentence anymore, until you find another sentence to put it with.”
With that, I had to end it so the kids could get to the library on time. Their teacher made sure they had all filled out their read-along worksheet that I’d prepared for them (you can find it here, along with the script I had kind of planned to follow), and then wrote on the board an actual fragment she’d read on a student test recently:
Yes, because of how dangerous it is.
Their homework? Do the kind of phrase diagram of it that we’d done with the complete sentence in class. Yikes — a noun clause with a fronted wh-phrase consisting of a wh-degree marker and an adjective, embedded under a compound preposition! I’d have liked to hear the discussion that generated the next day.