Faithful readers will know that my ideal alternative career is to be an editor in the Max Perkins mold. If not that, I think I’d enjoy being a literary essayist, someone like Alfred Kazin or Edmund Wilson or Louis Menand, who could write about my favorite authors and books in a forum where others would read and discuss what I wrote. I could occasionally collect my articles into books, and so on. On the other hand, if I actually had such a career, I wouldn’t have much of an option to do statistical research in my spare time, so I think for my own broader goals, I’ve gotten hold of the right side of the stick.
As it is, I enjoy writing about literary matters but it never quite seems worth spending the time to do it right. (And, stepping outside myself, I realize that I have a lot more to offer the world as a statistician than literary critic. Criticism is like musicianship–it can be hard to do, and it’s impressive when done well, but a lot of people can do it. Literary criticism is not like statistics. The supply of qualified critics vastly exceeds demand. Nobody is going to pay me $x/hour to be a literary consultant (for good reason, I’m sure), for any positive value of x. 9999 readers, Aleks,…
So you get it for free.
Anyway, this is all preamble to a comment on Clive James, who I just love–yes, I realize this marks me as a middlebrow American Anglophile. Deal with it. In any case, I came across this footnote in his verse collection:
Noam Chomsky gave furiously sleep ideas green colorless as an example of a random sequence of words which could have no meaning.
No, no, NO!!! This is so wrong that I’m wondering if James was making some sort of joke. But I can’t see what that would be. So, to straighten things out:
This is not quite correct.
1. He got the phrase backward: It’s “colorless green ideas sleep furiously.”
Chomsky used two examples: “colorless green ideas sleep furiously” and “furiously sleep ideas green colorless.” The former set of words sounds like a sentence (even though it makes no sense), the latter does not sound like a sentence (and also does not make sense).
2. It’s not a random sequence of words. It’s a very deliberate sequence of words, the reverse of a prhase that make perfect grammatical (or syntactic, I can never get these straight) sense. To use an analogy that James must be familiar with, “colorless green ideas” is like Jabberwocky–it sounds like English–all the parts of speech are in the right place. The difference, what makes the Chomsky sentence special, is that, first, the sentence makes no sense. But, beyond that, any two successive words of the sentence make no sense: Something green cannot be colorless, an idea cannot be green, ideas do not sleep, and you cannot sleep furiously. Chomsky’s sentence is a work of beauty, and it was disappointing to see Clive James miss the pointpartly miss this point–in a book of poetry, no less!
Just a couple words about Clive James. One thing that I find appealing about him is that he’s a writer in the David Owen mode (that is, David Owen the American journalist, not David Owen the English politician): serious, earnest, somewhat intelligent but a bit of a blockhead. Which I mean in a good way. Not clever-clever or even clever, but he wants to get things right.
P.S. Thanks to commenters for pointing out that my original blog was mistaken: Chomsky actually had two strings of words, not just one. in his famous article. So James did not get the phrase wrong (although he was in error in calling it “random”).
P.P.S. Yes, I realize that James is originally from Australia. Nonetheless, I think my enjoyment of his writing is more a sign of Anglophilia than Australiophilia on my part.
>this marks me as a middlebrow American Anglophile
Well he has lived there for quite a long time, but we Australians still claim Clive as one of our own….
Chomsky did use "furiously sleep ideas green colorless" as an example, but in that case to make the point that it was not grammatical (unlike "colorless green ideas sleep furiously").
Actually, Chomsky used both examples to make his points.
1. Neither sentence was likely to have ever been stated before.
2. A statistical test would find each one equally far removed from the english corpus.
3. But one sentence appears to have a grammatical structure.
How could this be? Chomsky's answer: deep structure.
Actually, James is right. Chomsky notes that in one order the words make no sense because of their syntax, and in the other order they make no sense because of their semantics. See also http://en.wikipedia.org/wiki/Colorless_green_idea… .
There's some statistics, or information theory, to be found in Chomsky's point as well. It's generally accepted that we all have some mechanism by which we can produce judgments like "Yes, that's a sentence," or "No, that's not a sentence." With colorless-green-ideas, Chomsky was directly arguing with researchers who thought this mechanism was driven by some kind of statistical model of language, based on n-grams or entropy or something (I'm really no expert here). Chomsky argued that the two strings
1) Colorless green ideas sleep furiously.
2) Furiously sleep ideas green colorless.
are both statistically improbable, but we know that 1) is a sentence, and 2) is not. Therefore, our grammatical judgment mechanism cannot be purely probability driven.
However, some statistical models can apparently distinguish the two, and say that 1) is more probable than 2).
http://en.wikipedia.org/wiki/Colorless_green_idea…
I don't have access to the James original at the moment, but I think he actually got it right. In Syntactic Structures (1957), Chomsky introduced the examples
(1) "colorless green ideas sleep furiously"
(2) "furiously sleep ideas green colorless"
to illustrate that semantically empty sentences could be both grammatical (the former) as well as ungrammatical (the latter):
"It is fair to assume that neither sentence (1) nor (2) (nor indeed any part of these sentences) has ever occurred in an English discourse. Hence, in any statistical model for grammaticalness, these sentences will be ruled out on identical grounds as equally "remote" from English. Yet (1), though nonsensical, is grammatical, while (2) is not grammatical." (1957:15)
Chomsky used both sentences in Syntactic Structures:
1) Colorless green ideas sleep furiously.
2) Furiously sleep ideas green colorless,
but then pointed out only the first one is actually grammatically correct, but both are (equally?) nonsensical.
Source:
http://tinyurl.com/j9jdd (wiki)
I think James has it right; Chomsky offered both the forward and backward directions — "colorless green …" with the description you gave, and "furiously sleep …" with roughly the description James gave.
It can be hard to do, and it's impressive when done well, but a lot of people can do it.
Do you think that's true? Or is it just that the demand for statisticians is far greater than that for literary critics?
Clive James is an Aussie not a Brit so that makes you an Aussie-o-phile. James is probably an Anglophile since he has lived there for decades. Unless, of course, liking an Anglophile makes you one a well, I suppose.
"Blockhead", in my experience, is generally interpreted in a pretty bad way. So if you are in a pub in London, I would not recommend calling someone that.
Chomsky's sentence is even better than you state, because it's not just that it is syntactically correct but doesn't mean anything, NO TWO CONSECUTIVE WORDS mean anything, and he asserted or implied, probably correctly, that none of these word pairs had ever previously occurred together in a sentence: "Colorless green", "green ideas", "ideas sleep", "sleep furiously." Any two words are nonsense. And yet, we have no problem identifying the subject of the sentence (ideas), figuring out what the ideas do (they sleep), figuring out what kind of ideas they are (colorless green ones), or describing how they sleep (furiously). This came up in a discussion of how people learn language — or, rather, to debunk a theory of how people learn language — of which I've now forgotten the details. But I agree, Andrew, it seems almost incredible that someone could have heard of the sentence, and could even remember the words in it, without getting the point of it at all.
To all:
(1) Thanks for the correction. I hadn't known that Chomsky had done it both ways.
(2) Still, Chomsky is not quite right. He describes "furiously sleep ideas green colorless" as a "random sequence of words." But it's not random at all.
To Phil:
You should read my blog more carefully before commenting. Above, I wrote, "any two successive words of the sentence make no sense." This is pretty much what you wrote above, except that I used italics and you used all caps.
I like Clive James too. One of the things he is best known for (at least amongst Brits of my age) is "Clive James on TV" where he would host a TV show, comprised of clips from TV shows from other countries, and mock them. Americans probably won't be familiar with that.
"Grammar" and "syntax" refer to the rules that govern the structure of language. Grammar is NOT meaning. Grammar is NOT in the dictionary. If you think of grammar as being things like "do not end a sentence with a preposition" or other 8th grade lessons — be they overly prescriptive or not — then you're on the right right track. It's rules for building sentences, basically. (We learn this stuff without thinking, and learn it early in our lives.)
Semantics is about meaning. The meaning of words, of phrases, of sentences, etc.. People decry "arguing about semantics" they are decrying little arguments about the exact meanings of words or details. Semantics are VERY local and particular (i.e. what does THIS sentence mean).
So, we look at the forwards example and it looks/sounds like english. It's following all the rules for english, with the right kinds of words in the right places, properly conjugated and agreed and everything. That's all grammar/syntax. Grammatically, no problem. The backwards example, though, does NOT make gramatical sense to English speakers, though it is easy to imagine an alternative language who's sytax was like English, just with the opposite word word order. You are right, it is NOT random. It follows rules, and therefore follows a grammar — just not quite ours.
But when we look at the forwards example, there is no meaning there for us. Semantically, it does not make sense.
This is well trod ground, and Andrew's right.
Any decent model of what sentences are likely to be spoken will take into account that it's more likely to have a noun following an adjective than the other way around. If only Mosteller and Wallace (or Zelig Harris or Yehoshua Bar-Hillel or Warren Weaver before that) had had more computer power, it'd be a different linguistics world today. Their footnotes thanking students for working out computations on slide rules and index cards are charming, but show how limited practical statistical study of language was in the early 1960s.
Here, as elsewhere in his work, Chomsky argues from lack of imagination. Because he doesn't understand how something could work or can't prove a theorem, he declares it impossible or false. Other examples include his dismissal of context-free grammars because they couldn't handle unbounded dependencies (they can), and the whole nativist view of language where he can't imagine how we could learn language wihtout a finite set of choices wired into our "language faculty".
Thank you, Bob Carpenter! Chomsky gets my hackles up all the time. His argument, often, seems little different from "God of the gaps" where God = the Language Acquisition Device.
Did I really miss that you said the two-words thing already? C'mon, you added it later, didn't you?
Phil: Nope, it was there all along.
It's funny, because you hold back posts for moderation you seem to get cascades of corrections when there's a mistake.
It also isn't that hard a sentence to assign meaning to, it's just a bit poetic.
"Ideas sleep" are just the latent bits of our zeitgeist not in active collective discussion, but somewhere in the collective unconscious waiting to rise back into active debate.
"sleep furiously" personifies these somewhat forgotten ideas, and expresses that they aren't happy about being shut out. It's that quiet seething before an explosive reoccurence.
"green ideas" may have made less sense when Chomsky first said it, but now that just means these ideas concern environmental stewardship.
"colorless" can be literal, but it can refer to something dull and pallid. So a colorless idea is a bland, uninteresting one.
So "colorless green ideas sleep furiously" is dismissing the banality of the environmental movement while worrying that it will have a violent resurgence sometime in the future.
Never underestimate the mind's ability to create logic where there is none :)
Paul: another parsing: Green ideas sleep furiously, unsettling the unconscious mind until they ripen into consciousness. Meaning is context — but I cannot make colorless work. So I defer to your ability to find meaning where the author meant none. Thank you Schleiermacher, Ingarden, and Gadamer.
Chomsky's sentence also is a vivid refutation of Skinner's operant language learning theory. As was suggested above, the likelihood of anyone having been exposed to that string of words in that sequence is low, but yet it is a syntactically meaningful sentence. Therefore, language is not exclusively learned by the shaping (reinforcement) of new language learners' utterances, but by learners developing an understanding of the deep structure of the language. If the only means by which we generated sentences was based on our reinforcement histories, we could not generate the colorless green idea, much less describe its sleep.