A new blog, The Diacritics (“thoughts on words”), looks to be worth reading; its creators say: “Hello! We’re John and Sandeep, two law students who think language is awesome. We write about how we use language in daily life, recent socio- and psycholinguistics research, and ways we see language at work in current events.” Recent posts include Why do some languages sound so fast? and Why are humans smart? Language and LEGOs. Welcome to the blogosphere, and keep posting!
Thanks so much for posting about our blog!
It took me an embarrassingly long time to correctly parse the sentence from the second article: “The girl the cake the baker the owner fired baked hit screamed”. I still haven’t clicked on the link but I’m sure I have it right now. At first I was fixated on the baker doing the hitting, rather than the cake (and presumably the baker doing the throwing). I wonder if this type of nested construction is any clearer in nominative-accusative or ergative-absolutive languages. In fact, is this construction possible in other languages at all? Examples, please…
If you insert “that” (or “which” or “who” or “whom”, according to taste) after each of “girl”, “cake”, and “baker”, maybe it becomes an example of something that’s possible in other languages, but still a little hard to make out the first time through.
It’s perfectly grammatical in English, it’s just not comprehensible because of human processing limitations. The same would be true in any other language; we can handle center-embedding to depth 2 or perhaps 3, but 4 is beyond us.
Would there be a British English equivalent of the Buffalo buffalo etc. sentence example because I’ve never heard of “buffalo” as a verb “to bully” before and would be more familiar with “to badger” or “to bulldozer” so I’m assuming the verb “buffalo” is more American English which of course I could be wrong about?
Maybe there’s something you could cobble together based on the different meanings of “March” for a Brit. Eng. version – March (the month); march (the noun and verb) March (the small town in Cambridgeshire)?
This native AmE speaker has never heard the verb ‘buffalo’ in any context other than that famous parlor trick sentence.
Well, it seems the attribution to Leftpondia (and Northicia) is correct. Here’s what OED2 says s.v. buffalo v.:
N. Amer. slang.
trans. To overpower, overawe, or constrain by superior force or influence; to outwit, perplex.
1903 Cincinnati Enquirer 9 May 13/1 Buffaloed — Bluffed.
1904 N.Y. Evening Post 25 Oct. 10 All the rest [of the newspapers] were what we used to term in the Southwest ‘buffaloed’ by the McKinley myth — that is, silenced by the fear of incurring the resentment of a people taught to regard McKinley as a saint.
1910 W. M. Raine Bucky O’Connor 77 O’Connor admitted that he was ‘buffaloed’ when he attempted an analysis of his unusual feeling.
1947 E. A. McCourt Flaming Hour 118 Jerry Potts himself would have been buffaloed.
I was sure I’d come across buffalo as a verb before the famous trick sentence but had trouble forming examples that sounded natural.
The last two examples given here finally clarified why. In my idiolect it’s either mostly restricted to agentless passives and/or buffaloed is an adjective restricted to predicates.
Being buffaloed into something seems kind of natural to me (not sure if I would say it but it’s not weird) while X buffaloed Y is definitely weird for me.
I think my idiolect is like michael’s in that regard.
Yes. X has Y buffaloed seems somewhat more natural (though it’s not common enough for me to say anything for certain).
At the risk of talking to myself while flogging a dead horse…may I repeat my plea for translations of “the girl the cake the baker the owner fired baked hit screamed” into other languages than English, keeping as near as possible to the recursive (nested) construction (causal chain proceeding from inner to outer)? Am I the only one here who thinks this is interesting? The article seems to regard this nested construction as a very clever and significant development in language. I don’t see how this makes any language without this feature inferior in any way though (not that the article says that it does).
“the girl the cake the baker the owner fired baked hit screamed
I’ll try some for Polish (caveat: I’m not a native speaker).
Okay, to start off with a possible rendering without the strict center embedding.
Dziewczyna, ktorą uderzyło ciasto, które upiekł piekarz, którego właściciel zwolnił krzyczeła.
The girl, who hit the cake, which baked the baker, who the owner fired, cried.
Polish doesn’t allow dropping of relative pronouns so they’re all there. And since they agree in gender, number and case with their reference (as do some verbs) the above isn’t hard at all to undertand. The word order in the relative clauses isn’t SVO all the time but again it’s pretty clear.
Trying to stay closer to the original.
Dziewczyna, którą, ciasto które, piekarz, którego właściel zwolnił, upiekł, uderzyło, krzyczeła.
Girl, who, cake, which baker, who owner fired, baked, hit, schreamed.
Now that’s awful and I’m sure no Polish speaker would accept it as well formed (whether for grammatical or stylistic reasons or both I’m not sure).
Maybe with participles (as close as Polish gets to dropping the relative pronoun).
Dziewczyna, uderzona przez ciasto, upieczone przez piekarza zwolnionego przez własciciela, krzyczeła.
Girl, hit by cake, baked by baker fired by owner cried.
No. Doesn’t work.
Dziewczyna, przez ciasto, które piekarz, przez właściciela zwolniony upiekł uderzona krzyczeła
Girl, by cake, which baker, by owner fired, baked, hit, cried.
No. Doesn’t work.
So provisionally, Polish doesn’t allow the structure of the original though attempts to translate it reveal some nice features that English lacks.
@michael farris: Thank you for that. You have made my day!
>Dziewczyna, ktorą uderzyło ciasto, które upiekł >piekarz, którego właściciel zwolnił krzyczeła.
>The girl, who hit the cake, which baked the >baker, who the owner fired, cried.
I presume you mean, as a more grammatical English translation: “The girl, who(m) the cake hit, which the baker baked, who(m) the owner fired, cried.” In my opinion, this is a little awkward in English because the “which” and second who(m) do not immediately follow the corresponding noun. I presume this is avoided in the Polish by the switch in the order from NV to VN. And presumably this is still perfectly OK in Polish because nominative and accusative cases are distinguishable. Hence the Polish is quite grammatical. Is my interpretation correct?
“Is my interpretation correct?”
First I notice a mistake, it should be krzyczała and not krzyczeła, not sure how that happened….
But yes. In Polish (unlike Hungarian and I think German) the relative pronoun needs to be next to the noun it applies to but word order within the relative clause can be arranged to make that possible.
Actually you could begin the whole thing with krzyczała
“Krzyczała dziewczyna, którą…..”
The weirdness of the English example seems to be an artifact of relative pronoun dropping and the need to keep the subject before the verb in each case. Languages without either of those two features are less liable to construct that particular kind of weirdness.
Again, there might other mistakes in my example and there’s no guarantee that a native speaker would accept my version(s).
Das Mädchen, das den Kuchen, den der Bäcker, den der Eigentümer feuerte, buk, schlug, schrie.
Why does the girl hit a cake??? Anyway, the relative pronouns cannot be dropped in such constructions, and they distinguish nominative and accusative (in the masculine), so the only thing that limits understanding is the chain of verbs. More than two verbs in the same person & number in a row aren’t comprehensible.
With a little dummy subject, you can fake that in German: Es schrie das Mädchen, das den Kuchen, den der Bäcker, den der Eigentümer feuerte, buk, schlug. I can’t tell anymore if that’s comprehensible… probably not. It’s poetic, though (because of the dummy subject).
“The girl the cake the baker the owner fired baked hit screamed” Does everyone agree that this is “perfectly grammatical English”? It doesn’t seem so to me. I guess I’d always figured that the rule that allows you to omit the relative pronoun only applies on one level of nesting.
Komfo:
We know it’s a processing problem rather than a grammatical error because intelligibility depends on the specific lexical choices. Consider these four:
1) The horse the man rode galloped.
2) ?The dog the dog bit bit.
3) ??The street the horse the man rode galloped on is long.
4) *The dog the dog the dog bit bit bit.
#1 is no problem; #2 is questionable — naive listeners would find it hard to understand; #3 may barely be processable; #4 is complete gibberish. So even though #1 and #2 are both singly center-embedded, and #3 and #4 are both doubly so, #2 is significantly worse than #1, and #4 is infinitely worse than #3.
@David Marjanović:
>Das Mädchen, das den Kuchen, den der Bäcker, den >der Eigentümer feuerte, buk, schlug, schrie.
>Why does the girl hit a cake???
Is that what the German means? Then it is a mistranslation. In English “The girl the cake hit screamed” means the cake hit the girl, then the girl screamed. Isn’t that what “Das Mädchen, das den Kuchen schlug, schrie” means too?
Now I think I see what’s going on. My internet research reveals that the article in “den Kuchen” is in the accusative and so indicates that cake is the object of hit, rather than subject. So the correct German translation should be “Das Mädchen, das der Kuchen, den der Bäcker, den der Eigentümer feuerte, buk, schlug, schrie.” And Mädchen (girl) has neuter gender. How weird. Am I right?
Huh. *blink* That makes even less sense.
Yes.
What’s weird is that all diminutives automatically have neuter gender in German. This one is derived from Magd (today “female farmhand”, in earlier times “young woman”), which has feminine gender just as you’d expect.
I have continued thinking about this (maybe a little too much). It seems to me the simplest way of expressing the sentence in question (in English and probably all languages?!) is to break it up (hmm…breaking it up & breaking it down means the same thing!) into separate short Hemingway-esque sentences e.g. “The owner fired the baker. The baker baked the cake. The baker threw the cake. The cake hit the girl. The girl screamed.” I have the idea that some languages are averse to long sentences with lots of clauses and sub-clauses, and prefer short multiple sentences. I am thinking of Chinese for example. Is this just my ignorant prejudice or is there something in it?
The simplest way to express the whole episode in a single sentence would be what I would call a linear active structure e.g. “The owner fired the baker who baked a cake that he threw at and hit the girl who then screamed.” where the causal chain proceeds from original cause to ultimate result. Maybe most (all?) languages (including Chinese?) can do this, stringing out the sentence indefinitely?
The next simplest way might be what I would call a linear passive structure, such as “A scream was let out by the girl who was hit by the cake which was baked and thrown by the baker who was fired by the owner.” In this case the causal chain s reversed from ultimate result to original cause.
Then, more complicatedly, there could be a mixture of active and passive. With four subclauses (as here) that means a total of 2x2x2x2=16 ways of expression, including the pure active and pure passive just referred to.
For example, there are six possible sentences with 2 active and 2 passive subclauses:
1. The girl the cake baked by the baker fired by the owner hit screamed.
2. The girl hit by the cake baked by the baker the owner fired screamed.
3. The girl hit by the cake the baker fired by the owner baked screamed.
4. The scream was let out by the girl hit by the cake the baker the owner fired baked.
5. The scream was let out by the girl the cake baked by the baker the owner fired hit.
6. The scream was let out by the girl the cake the baker fired by the owner baked hit.
Then, most complicated of all is the totally nested structure which can even be expressed without articles in a certain dialect of English (where tense is usually “present)” e.g.
“GIRL CAKE BAKER OWNER FIRES BAKES HITS SCREAMS!”
It seems to me from David Marjanović’s comments that the nested construction is equally grammatical in German as it is in English, not surprising perhaps since they both belong to the Germanic language family. What about Romance languages (say French, Italian or Spanish)? Of course relative pronouns may be included if they are required. Without further evidence, I may form the tentative hypothesis that it is a purely Germanic thing.
The periodic style of, say, Aristotle and Cicero leads to center embedding.
That was missing from the example!!! With it, everything suddenly makes sense. Replace schlug “beat” by traf “hit [a goal]” in the German version, and put warf “threw” in somewhere.
Northwest Caucasian sentences are said not to allow dependent clauses at all, except for “limited” ones in Abkhaz that are blamed on Russian influence. Alas, Wikipedia does not provide examples.
Chinese is one of those that doesn’t do it in the Standard Average European way. Instead of relative clauses, it uses the particle de which turns what precedes it into an attribute. “A Chinese teacher is a teacher who teaches Chinese” = “a Chinese teacher is a teacher of teaching Chinese” = Hànyǔ lǎoshī shi jiào Hànyǔ de lǎoshī.
Actually, I think it works at least in French.
@MMcM: Thanks for that. I didn’t know about Aristotle, Cicero and periodic sentences. I have some reading to do.
@David Marjanović: I had an uneasy feeling that the ambiguity of the English “hit” might have caused some confusion. I am interested to learn about the disambiguation provided by the German “traf” and “schlug”.
“Если на клетке слона прочтёшь надпись «буйвол», не верь глазам своим.” -Козьма Прутков.
[If you read “buffalo” on an elephant’s cage, don’t believe your eyes. -Kozma Prutkov]
Treffen, simple past traf, is also “meet”.