Home             Musings

On Frobnication, Universal Grammers,
and the Pela Bilong Missus Kwin

A review of The Language Instinct by Steven Pinker

This is a talk I gave one Friday in 1995 at Mount Stromlo Observatory as a diversion from astronomy, and several observatory members make cameo appearances here. In particular, Brian the bringer of bagels is the distinguished Brian Schmidt and he gave the following talk, which was an introduction to wine tasting.

This is actually quite a serious talk. It all began a few Fridays ago when Ron was talking about velocity fields, and used the word frobnicate to mean fit (as in fit parameters). Like most of us, I'd never heard of frobnicating, but I'd just been reading Steven Pinker's The Language Instinct, and knew all about frobbing, and so I thought I knew exactly what Ron meant. But actually he didn't mean what I thought he meant....Anyway, with the raging discussion about frob and frobnicate that went on after Ron's talk---at one point Mike Bessell asked rhetorically ``What language are you talking about?''---I thought it would be interesting to come up here and review the book. After all, language is something that concerns us all. And the purpose of this talk is not to summarize this book---which I can't---but to persuade you that it's worth your while to read it for yourself.

Steven Pinker is a linguist from the school of Noam Chomsky---every discussion about languages these days seem to get to Chomsky eventually. This book is sort of about how languages work, and what is truly amazing is that there seems to be a method behind the madness of the thousands of languages that humans speak. People have known about similarities in vocabularies for a long time. For instance English and my mother tongue, Bengali, are at opposite ends of the family of Indo-European languages which includes several Asian ones and nearly all the European ones. What people have learned since the 1950s---and I gather this is mostly due to Chomsky---is that there are fundamental similarities in the grammars of all languages. There seems to one universal grammar with a few adjustable parameters that each language sets. Actually, each language will typically have several dialects, each with a slightly different parameter setting. Spoken and written language are never grammatically identical---if this talk sounds stilted, I apologize; trying to write it has confused my parameter settings a bit....

A rather telling study Pinker cites is a survey of the fraction of grammatical sentences in speech in different social classes and contexts. ``Grammatical'' here and forevermore means ``well-formed according to consistent rules in the dialect of the speakers'', and ``ungrammatical'' refers to random sentence fragments and general word salad.

The great majority of sentences were grammatical, especially in casual speech, with higher percentages of grammatical sentences in working-class speech than in middle-class speech. The highest percentage of ungrammatical sentences was found in the proceedings of learned academic conferences.

This is a long way from the self-styled defenders of the purity of language, and Pinker has a whole chapter to say rude things about the busybodies who tell you not to split infinitives, and not to end sentences with prepositions. It turns out that most of those rules about English are fads from the eighteenth century. When the British Empire was expanding, the London dialect was suddenly an important world language. And suddenly it had to be respectable. How to make English more respectable? Why, make it more like Latin of course! (What better way to get a big empire than to start talking like the Romans?) So you think of things you can't do in Latin, like split infinitives (because Latin infinitives are single words), and forbid them in English.

To go boldly where no man has gone before? Beam me up Scotty; there's no intelligent life down here.

So the message is: if the sound feels OK, you would be following some consistent rules, and you should stop worrying about it and start worrying about other things, like getting some meaning across.

But what are those rules? Well a book like this can obviously only go into some basics, and even that gets heavy-going in places, so I'll give a couple of examples. English is a subject-verb-object language; you say:

Brian brought bagels in a big basket.

So subject (Brian), verb (brought), object phrase (bagels in a big basket). Also, the preposition comes before the noun phrase it preposes to: `in a big basket'. Japanese, is a subject-object-verb language. It puts verbs after, and puts prepositions after too:

Brian a big basket in bagels brought.

A question like

Did you have a bagel?


You a bagel have did?

Actually, I made those examples up, and I don't know Japanese. How did I guess? Well, I just took apart a sentence in Bengali, and said ``Aha, I a subject-object-verb language of example know!'' and just used my Bengali parameter setting.

Actually, that trick will work only for fixed word order languages. Bengali has fixed word order and English mostly does, but I don't know about Japanese. English may be in a transition now to variable word order. Brian's bagels was originally Brian his bagels. But when you've tagged Brian with an 's the bagels can wander all over the place:

The bagels in the big basket over there are Brian's.

Anyway, the general idea is that parameter changes can make drastic but entirely predictable changes in grammar.

How do small children figure grammar so perfectly, long before they have a clue what arithmetic is (for example)? The radical position is that babies already instinctively know the universal grammar, all they have to do is to work out the parameter settings and memorize the word roots. So when baby hears ``Eat your spinach!'' (which doesn't even have a subject), baby thinks ``Eat came before spinach; hmmm....this is probably one of those subject-verb-object languages.'' But even you don't buy that scenario, you can hardly avoid concluding from the universalities that some component of human language is innate. Chomsky likes to say that a visiting Martian anthropologist would conclude that apart from some mutually unintelligible vocabularies, all humans speak the same language.

So what about vocabularies? Have a look at this list:

ptak thale hlad
plaft sram mgla
vlas flutch dnom
rtut toasp nyip

None of these are English words. Which ones could have been, though? All the phonemes here are found in English, but most of the syllables don't feel right, yet a few do. This shows that neither phonemes nor syllables are the most sensible basic units for English. The natural basic units are something in between, called onsets and rimes. A syllable is an onset followed by a rime. An onset is one of more consonants, and a rime is a vowel followed by one or more consonants. Only certain onsets and rimes appear in English; for example `sr' is pronounceable, but not legal in English. But there's no formula for predicting which onsets and rimes will be legal. But there is a formula for predicting which syllables feel Englishy---you just combine a legal onset with a legal rime.

Frob is definitely Englishy. It's hacker slang for `twiddle aimlessly'. The etymology (thank you Kim and Dave for tracing this) seems to come from the proper noun Frobnitz via frobnicate. But while frobnicate isn't very interesting to anyone past adolescence, frob is very interesting. If you try saying Miao, and Maui, you'll find that you can say Miao when you're nearly asleep, but Maui is harder. The order of vowels Miao make it one flowing gesture for the mouth, but Maui is more complicated. And you find this easy vowel sequence again and again in English: riff-raff, tic-tac-toe, bibbity-bobbity-boo. Across languages, there's a noticeable tendency for the earlier vowels to associate with little things, and the later vowels with big things. Frob fits this pattern quite nicely, and it fills a real need for a word. When you're trying to fit parameters, you first frob, and then you tweak. This is just the sort of thing that would come into general usage. And when someone says they frobbed the parameters, I understand that they did a very coarse fit. That's why I got confused by Ron, because he actually didn't mean frob, he meant tweak....

Well, I've already used frobbed, and frobs and frobbing sound fine, so clearly this isn't going to be an irregular verb. In fact, irregular verbs and nouns are a nearly closed club, and seem to be relics of the lost Indo-European proto-language. Irregulars have quite a hold on us, though. Here's a fun experiment with 3 year olds. You show them a picture of a monster and say ``This is a monster that likes to eat mud---we call it a mud-eater.'' Another picture and ``This monster likes to eat mice---what do we call it?''. You get the instant response mice-eater. Another picture and ``This monster likes to eat rats---it's a?''. Rat-eater, notice not rats-eater. But mice-eater, not mouse-eater. Irregular nouns get pluralized, regulars don't. But the situation changes if the compound has a completely unrelated meaning; then compounds are always regular. Thus a workman is a kind of man, and the plural is workmen; but a low-life is not a kind of life, and the plural is regularized: low-lifes, not low-lives. Walkman is a funny one: Walkmen sounds awful, but Walkmans doesn't feel quite right either. Pinker suggests that's because it's not a standard English compound, but a Japanese compound from English words.

And on the subject of new English words, we can't go by without the first stanza of Jabberwocky:

'Twas brillig, and the slithy toves
Did gyre and gimble in the wabe;
All mimsy were the borogoves,
And the mome raths outgrabe.

Lewis Caroll makes up all sorts of new verbs, nouns and adjectives. There's even an irregular verb in there---as you know the past tense for outgribe is not outgribed but outgrabe. But he doesn't touch the function words---prepositions, articles and so on. In fact, function words in a language are really resistant to change. That's why new gender-neutral pronouns just haven't worked in English. When we want a gender-neutral pronoun, we tend to commandeer an existing one and overload it. Like using they as a singular pronoun, which is quite standard now.

If it's that easy for new words to appear, what about new languages? New languages appear all the time too, and Pinker gives us several examples.

When a group of adults without a common language gets together they tend to pool words from different sources and make up a pidgin. That can happen among immigrants, as in Hawaii in the last century, or when long isolated communities meet, as in Papua New Guinea. Pidgins don't have much grammatical structure, and that leaves ambiguities in speech which have to be resolved from context or other cues. Children obviously don't like pidgins, because the moment they hear one they put a grammar on it and turn it into a creole. And here's the one blooper I spotted in the book. Pinker gives Papua New Guinean Pidgin as an example of a pidgin, where Prince Philip is known as `Pela bilong missus Kwin'. In fact, by his definition it's not a pidgin at all, but a creole. `Pela bilong missus Kwin' is no random collection of English words; it's a perfectly grammatical phrase which just happens to have English word roots. From other quotes I've seen, I'm pretty sure bilong is not a verb but a preposition. That's a giveaway that this really is a new language---remember we said prepositions are a closed club. And bilong's position tells us this is another subject-verb-object language. And I suspect missus is not a title but a noun classifier. English has only a few of those, like `a gaggle of geese', but some languages use many.

Then there are artificial languages. There's Esperanto, which is fashionable to put down these days, but it has a goodly number of speakers, so it must be doing something right. A really extraordinary one is Damin, which the Lardil Aboriginals invented. It was a ceremonial version of their everyday language, but they transformed it in two way. First, they pared down and altered the vocabulary, so it could ``express the full range of concepts in everyday speech'' (whatever that means) with about 200 words. The other thing they did was to introduce a whole lot of new consonants: some clicks which you also find in southern Africa, and some consonants you find nowhere else. The one word I know, is for fish (bony fish, actually) and it has one of those weird consonants: it's [L]ee. Why did they do this?

Have you ever seen a deaf language? Yes, the deaf have their own languages, and have always had them. The most studied one is American Sign Language, or Ameslan. It's completely unrelated to English or any spoken language, but descended from a sign language which deaf down-and-outs in France have had for centuries. There's Ozlan in Australia, but I don't know anything about it, except that it looks very beautiful. By now you won't be surprised when I tell you that sign languages have grammars, and can be classified just like spoken languages. (Ameslan turns out to be a highly inflected language, which I guess means it has a very fluid word order.) Pinker gives an example of a pidgin and a creole sign language too. And you know what babies of babbling age do when they're exposed to sign language?

I'm nearly done now, but you can't borrow my book yet, because Tracey's borrowing it first. But if you do read it, I think you'll come away feeling a little freer than you did before. To show what I mean, I'll end by quoting a quote from Mark Twain, though not for quite the same reason as Pinker quotes it. It's part of a translation into English (by Mark Twain) of an address he gave to a Vienna Press Club. The original was in German; Mark Twain spoke excellent German---he'd have had to, to get away with this. Once again, this is someone doing silly things with the language, and once again Mike might ask ``What language is this?'' But if you forget about English grammar and think only about clarity and just general good writing then this---and such is the genius of Mark Twain---this is an exemplary use of English.

I am indeed the truest friend of the German language---and not only now, but from long since---yes, before twenty years already.... I would only some changes effect. I would only the language method---the luxurious, elaborate construction compress, the eternal parenthesis suppress, do away with, annihilate; the introduction of more than thirteen subjects in one sentence forbid; the verb so far pull to the front that one it without a telescope discover can. With one word, my gentlemen, would your beloved language simplify so that, my gentlemen, when you her for prayer need, One her yonder-up understands.

....I might gladly the separable verb a little bit reform. I might none do let what Schiller did: he has the whole history of the Thirty Years' War between the two members of a separate verb inpushed. That has even Germany itself aroused, and one has Schiller the permission refused the History of the Hundred Years's War to compose---God be it thanked! After all these reforms established be will, will the German language the noblest and prettiest on the world be.