Help us with a syntax question (buffalo buffalo buffalo.)

March 5, 2008 6:15 PM

Do fish fish fish fish fish fish fish?

So my colleague and I have come at odds about syntax.

He insists that the sentence "fish fish fish fish fish fish fish" is syntactically correct, while I'm pretty sure that he needs a participle or a demonstrative in there to make it work.

(A quick map of the argument)

Fish (direct object) fish (subject) fish (verb) fish (verb) fish (direct object) fish (subject) fish (verb).

...in other words, the fish that fish take sport in fishing also take sport in catching fish that the aforementioned fish take sport in catching.

However, I'm pretty sure that it should be written "fish that fish fish fish fish-fished fish" (or something like that.)


What say you, o mighty hive mind?

posted by ThomThomThomThom to writing & language (35 comments total)

This looks to be a variation of the famous "Buffalo buffalo buffalo...".
posted by kingjoeshmoe at 6:18 PM on March 5, 2008

You don't need that. "The man I love" = "The man that I love."
posted by languagehat at 6:20 PM on March 5, 2008

It works with the same structure as "Fish sharks eat, eat fish sharks eat" except without the comma and with fish fishing fish instead of sharks eating them.
posted by nicwolff at 6:24 PM on March 5, 2008 [3 favorites]

It looks like the end of the Buffalo Wikipedia entry says it works with fish.
posted by 4ster at 6:30 PM on March 5, 2008

It makes sense if you read the first two fish to mean "fish that are uber-fish." Sort of like, a MOVIE move or a KISS kiss. ("That wasn't just a kiss; that was a KISS kiss!")

Do FISH fish do something:

Do they fish for something in particular?

Do FISH fish fish for something in particular?

Do fish fish fish fish fish fish fish?

Do they fish for those fish who are also FISH fish?

Fish who are FISH fish are FISH-fish fish.

Do FISH fish fish FISH-fish fish? (Do Fish fish go fishing for FISH-fish fish?)
posted by grumblebee at 6:36 PM on March 5, 2008

Hmm I think it makes sense if the first "fish" are the opposite of uber-fish, like second on the fish food chain or something. Like, I'm a second class fish that other fish like to fish. My buddy is also a second class fish that other fish like to fish. What's more, I am not above fishing my second-class buddies.
posted by Pigpen at 6:57 PM on March 5, 2008

What's more, I am not above fishing my second-class buddies.

But that's okay, because they're not above fishing me either.
posted by Pigpen at 6:58 PM on March 5, 2008

He insists that the sentence "fish fish fish fish fish fish fish" is syntactically correct, while I'm pretty sure that he needs a participle or a demonstrative in there to make it work.

This is "grammatical" in the sense linguists use the word. A demonstrative (actually, "that" is not a demonstrative here, it's what is called a complementizer) isn't needed because this is what's called a reduced relative clause (this is what languagehat's example demonstrates). In English it is typically ok to leave out the complementizer in this construction. A participle isn't needed either; it could be used to get the meaning "fish that are fished", but this is accomplished by "fish (that) fish fish".

However, and this is the important part, "grammatical" does not mean easy to process, or easy to understand. The sentence works because "fish" is ambiguous between a noun and a verb, and the reduced relative structure is a functionally useful construction because in many cases "that" is not needed to determine that you have a relative clause. However, here both factors that lead to its grammaticality make it very hard to parse. "Fish that fish fish fish fish that fish fish", with non-reduced relative clauses is much easier to process. "Fish that fish fish, they fish fish that fish fish" with what is called a "left dislocation" structure is even easier, because the left dislocation makes it clearer that the fourth "fish" is the main verb of the sentence (something that is usually obvious for many reasons at once). So this processing problem accounts for your intuition that the sentence is wrong in some way. That is, if you're trying to communicate by using this sentence, it is wrong.

The other canonical example of something that is (usually claimed to be) grammatical but difficult to parse is a "center embedding" structure, as in (this example is from David Beaver on Language Log) "Starlings linguists language loggers readers follow commented on the work of studied are damn smart!".
posted by advil at 7:02 PM on March 5, 2008 [3 favorites]

This is "grammatical" in the sense linguists use the word.

I disagree. Linguists use the word descriptively, not proscriptively. Now, sure, you could take such and such research saying that english syntax has been observed as behaving such-like, under which observations, the sentence given would fit under the norms of english speakers. But go show that sentence to 10, 100, 1 million people and see what they think. Now, I can't cite such a study, because it does not exist yet. But, my hypothesis is that very, very, very few of them would recognize it as a possible english sentence, and even fewer would interpret it as originally intended. Ultimately, you can try combining two observations in specific circumstances, and hypothesize that they'll still hold up, but quite often, you'll find a new "exception to the rule" (I use quotes to insinuate that there are not rules and exceptions to them, just ways people talk, and ways they don't, and then ways somewhere else hundreds of miles away, where they do talk the way they don't in the original locale, etc.), and then you have to go about studying that exception to describe it accurately.
posted by gauchodaspampas at 7:16 PM on March 5, 2008

There's another famous one.

"Oysters eat." Clear enough.

"Oysters oysters eat eat." That is, oysters that oysters eat, themselves eat.

Now... the following is said to be grammatical, but I find it incomprehensible:

"Oysters oysters oysters eat eat eat."

My brain just shuts down trying to figure out which oysters are eating what.
posted by kindall at 7:23 PM on March 5, 2008

Linguists imagine that they know the specific production rules for the English language, and that any sentence which can be derived from those production rules they've written down is in fact a valid sentence in the English language.

This misses the fact that the sentence in question is quite simply gibberish. It's a pile of words that a native English speaker knows is not a valid utterance in the English language.
posted by jepler at 7:25 PM on March 5, 2008 [2 favorites]

Other grammatical but hard to understand sentences include garden path sentences:

The horse raced past the barn fell.
The old man the boat.
The cotton clothing is made of folds easily.
posted by painquale at 7:27 PM on March 5, 2008 [2 favorites]

ThomThomThomThom says fish fish fish fish fish fish fish like Buffalo buffalo Buffalo buffalo buffalo buffalo Buffalo buffalo

eponysterical!
posted by seawallrunner at 7:29 PM on March 5, 2008

It works for me.
posted by M.C. Lo-Carb! at 7:32 PM on March 5, 2008

I've tried substituting unique words for each of the words in the original 'fish' utterance, and I still don't believe it's an English sentence without adding some 'that's:

Dogs that nip cats nip dogs that nip cats.

vs

Dogs nip cats nip dogs nip cats

I can believe the first is a sentence but not the other. And if I ever wanted to talk about that, I'd probably choose a different construction, like

Dogs that nip cats nip other cat-nipping dogs too

Just like Newtonian physics will sometimes fail to describe the actual behavior of real objects, so will the supposed "rules" of English occasionally fail to describe the actual usage of English by fluent speakers of the language.
posted by jepler at 7:32 PM on March 5, 2008

Linguists imagine that they know the specific production rules for the English language, and that any sentence which can be derived from those production rules they've written down is in fact a valid sentence in the English language.

That is exactly wrong. Linguists look at how people speak, and they describe it. They don't recognize right and wrong. They recognize people (sometimes much more specific than that, like middle class, college educated, native english speakers who frequent the site metafilter.com) speak like this, but people don't speak like this, unless of course they are upper class, male, and frequent fark, and then only 73% of those in the study did.
posted by gauchodaspampas at 7:36 PM on March 5, 2008 [2 favorites]

jepler: You've got the sentence structure slightly incorrect.

To rewrite your example- Dogs cats nip nip cats dogs nip.
posted by zamboni at 7:41 PM on March 5, 2008 [4 favorites]

gauchodaspampas: This is "grammatical" in the sense linguists use the word.

I disagree. Linguists use the word descriptively, not proscriptively.

This is not the case at all -- the descriptivist vs. proscriptivist debate has nothing to do with this. "Grammatical" for pretty much any linguist simply means syntactically well-formed. All linguists will admit that insofar as no one can understand these sentences easily, there's something wrong with them, but there are many ways for a sentence to be deviant: syntactically, semantically, pragmatically, etc.

Maybe the standard person on the street uses "grammatical" to just mean "OK sounding". That's fine; the descriptivist linguist won't chastise him. He's just using the word in a different, less technical sense. But that does not mean that linguists don't have their own internal standards for the use of the word "grammatical.
posted by painquale at 7:48 PM on March 5, 2008

I disagree. Linguists use the word descriptively, not proscriptively.

Painquale has it exactly right. By grammatical I mean syntactically well-formed (the vast majority of descriptivists do assume such a notion), I probably should have been more explicit.

They don't recognize right and wrong. They recognize people (sometimes much more specific than that, like middle class, college educated, native english speakers who frequent the site metafilter.com) speak like this, but people don't speak like this, unless of course they are upper class, male, and frequent fark, and then only 73% of those in the study did.

This may be true of small number of linguists who work in a very specific, very traditional, corpus based approach, and to a certain extent people doing research that is specifically on sociolinguistic factors in dialect variation. It is not an accurate description of the modern descriptivist tradition, which tries to analyze the commonalities between speakers' mental grammars, not simply describe who has what grammar and what the differences are.

Also, the idea that descriptivism means "no speech is wrong" is not correct. Descriptivist work fundamentally assumes that speakers have a "mental grammar" of some kind, and that there will be utterances that would be incorrect in terms of that grammar. It even assumes that speakers themselves are generally reliable at determining right from wrong. The difference from prescriptivism is that descriptivists are trying to figure out that grammar, not tell others how to speak. In fact, it is commonly recognized that speakers produce speech errors all the time (e.g. involuntary spoonerisms and so on); there are in fact many linguistics and psychologists who study speech errors, since despite being wrong, they can be revealing of how the grammar is organized. People have even collected databases of such errors.
posted by advil at 8:27 PM on March 5, 2008 [2 favorites]

Also there's a big difference in reading a sentence and hearing it. A lot of information can be conveyed in prosody. Anyone who remains unconvinced that this is a sentence of English: try to track down someone in real life who understands its syntactic structure and have him read the sentence out loud to you. You might change your mind.
posted by tractorfeed at 9:32 PM on March 5, 2008

Fish buffalo fly fish buffalo fish buffalo fly fish.

That is all.
posted by flabdablet at 1:09 AM on March 6, 2008

Though it is gramatically possible to leave out "that", it is possible to create more interesting effects than "buffalo" and "fish" sentences. Try the following sentences, for instance:

The fat people eat accumulates.
The horse walked past the barn fell.

It's an interesting abuse of syntax since the sentences are syntaxically correct up until the last word. Though the last word doesn't make the sentence incorrect, it forces you to parse again the sentence from the start.

The point is that subclauses are not required to be separated from the sentence with any words or commas.
posted by cotterpin at 4:50 AM on March 6, 2008

Now, sure, you could take such and such research saying that english syntax has been observed as behaving such-like, under which observations, the sentence given would fit under the norms of english speakers. But go show that sentence to 10, 100, 1 million people and see what they think. Now, I can't cite such a study, because it does not exist yet. But, my hypothesis is that very, very, very few of them would recognize it as a possible english sentence, and even fewer would interpret it as originally intended.

It's difficult to recognize when written. But I don't think it's difficult to recognize when spoken. Or, at least, not as difficult as you're imagining.

It's not like they're going to be monotonically saying, at constant speed, "FISHFISHFISHFISHFISHFISHFISH".

Several different tones would (or could) be involved, even different tones within a single "fish", each of which (and together) imply something about the sentence structure to a native listener. Along with appropriate vocal pauses, I actually think the sentence is fairly clear.
posted by Flunkie at 5:12 AM on March 6, 2008

Interestingly, German has something very similar, though it's slightly easier to parse since nouns are capitalized.
"Wenn Fliegen fliegen hinter Fliegen, fliegen fliegen Fliegen fliegen Fliegen hinterher."

The literal translation is "when flies fly behind flies, fly flying flies flying flies behind."
posted by belladonna at 6:10 AM on March 6, 2008

I disagree. Linguists use the word descriptively, not proscriptively.

Wow. Not only are you telling an actual linguist how linguists use words, you don't even know the difference between prescriptive and proscriptive. You fail.

Needless to say, advil is perfectly right, and the sentence is perfectly grammatical. And this is bullshit:

It's a pile of words that a native English speaker knows is not a valid utterance in the English language.

I'm a native speaker, and I know it to be a valid utterance. It just isn't immediately parsable on first reading. If that troubles you, stay away from literature.
posted by languagehat at 7:18 AM on March 6, 2008

Advil makes my head hurt.
posted by grateful at 8:21 AM on March 6, 2008

First of all, thanks to nicwolff for explaining this in a way I can finally get my head around--I've tried for a while to parse the buffalo sentence and never gotten it until now.

But it also occurs to me that the buffalo sentence in Wikipedia, with 8 occurences, is actually shorter than it could be. Consider that you can write a sentence with 7 buffalo referring only to the animal and the action, never to the city, using the same syntax as the fish sentence here:

Buffalo buffalo buffalo buffalo buffalo buffalo buffalo.
(Bisons [which] bisons intimidate, [themselves] intimidate bisons [which] bisons intimidate.)

Now you can take every incidence of "buffalo" meaning "bison" in that sentence, and put "Buffalo" the city in front of it, to get a sentence with eleven buffaloes:

Buffalo buffalo Buffalo buffalo buffalo buffalo Buffalo buffalo Buffalo buffalo buffalo.
(New York bisons [which] New York bisons intimidate, [themselves] intimidate New York bisons [which] New York bisons intimidate.)

Now if only Buffalo New York had a sports team called the Buffalo, we could get even more in there...
posted by DevilsAdvocate at 10:19 AM on March 6, 2008

Slightly off-topic, but perhaps you and your colleague would enjoy a song by Mr Scruff by the name of "Fish". Here it is on youtube - make sure you wait for the "chorus" at 01:49.
posted by urban greeting at 10:45 AM on March 6, 2008

"Oysters oysters oysters eat eat eat."
sounds like a cheer for a pie-eating team with an unfortunate mascot.
posted by prophetsearcher at 2:04 PM on March 6, 2008 [2 favorites]

Chicken chicken chicken (previously).
posted by attercoppe at 10:10 AM on March 7, 2008

CLAIM: the following is a grammatical sentence.

"BADGER BADGER BADGER BADGER BADGER BADGER BADGER BADGER BADGER BADGER BADGER BADGER MUSHROOM MUSHROOM."

Mushroom-noun-1: a type of fungus
Mushroom-adj-1: ephemeral, upstart

Badger-noun-1: a type of mammal of the Mustelidae family
Badger-transitiveverb-1: to trouble persistently
Badger-adj-1: Native to Wisconsin
Badger-adj-2: Especially resembling a badger
Badger-adj-3: member of the 1970's rock band formed by Tony Kaye after he left Yes

Now you can give it the same structure as the Buffalo sentence:

[B-adj-1 B-adj-2 B-adj-3 B-noun-1] [B-adj-1 B-adj-2 B-adj-3 B-noun-1] B-verb-1 [B-adj-1 B-adj-2 B-adj-3 M-adj-1 M-noun-1]
posted by painquale at 5:34 PM on March 9, 2008

Nice try, painquale, but I don't think that works. "Fish" and "buffalo" work because those two, as nouns, can be plural. The plural of "badger" as a noun is "badgers." So some of the instances of "badger" in your sentence would have to be "badgers" to be grammatical. If you were just doing a two-word, noun-verb sentence, "Fish fish" and "Buffalo buffalo" both work (plural noun, plural verb), but "Badger badger" does not--it would have to be "Badgers badger."
posted by DevilsAdvocate at 9:11 AM on March 10, 2008

Grr, you're right. I did give that some consideration, but decided that 'badger' could be used as a generic. But I guess that's not the case. "Chicken lay eggs" and "fox hunt geese" sounded OK to me, so I figured I was good to go. But 'chicken' is somewhat special (the OED says that "chicken sometimes occurs as a plural or collective."), and the 'fox' sentence is ungrammatical; I guess it sounded good to me because 'fox' ends in a fricative and so sounds like a plural.

I guess that if I want to save the grammaticality of the badger sentence, I have to claim that 'badger' can be treated as a plural, like 'chicken'. I dunno, maybe that's not unreasonable. If the voice-over in some documentary said something like, "Badger are found mainly in temperate environments...", I don't know if I'd notice that it sounded weird. OK, let me then claim that in my idiolect (and the idiolect of anyone else who finds the above sentence good), 'badger' can be a plural and the badger sentence is grammatical. (Well, actually, I have to do the same trick with 'mushroom', and that seems much less plausible to me.)
posted by painquale at 12:36 PM on March 10, 2008

It works fine if it's a command: "Badger, badger mushroom!"
posted by flabdablet at 3:25 PM on March 12, 2008

That's why the most used languages have irregular rules for commonly used phrasings (thank god that flower is holding back the it's its thing).
posted by porpoise at 9:43 PM on March 23, 2008