This sentence from Mark Liberman’s (very interesting) Language Log post on the history of sentence diagramming (which I, like Mark, had to learn as a lad) is a good example of why I dislike the culture of MIT linguistics (the linguistics itself is a whole other issue): “And there was not a great deal of respect for earlier traditions of analysis—the required ‘History of Linguistics’ course at MIT was familiarly known as ‘Bad Guys’.”
Here at LH, those pre-Chomsky linguists are known as the Good Guys.
I don’t understand this attraction to pre-Chomsky linguists, especially of the descriptive variety. Why should actual data get in the way of a perfectly good theory?
Not that the “bad guys” fare better even when their work wins out. Carl Pollard once wrote (sarcastically) of a proposal adopted by Chomsky without appropriate credit: “Why should Gerald Gazdar and Ron Kaplan get credit for having those ideas when they were wrong?”.
Um, not to start a linguist flamewar here, but I’m not sure the Chomskyans should be talking about letting data get in the way of theory, given the way they have played fast and loose with actual data from actual languages…
Hm? I’m not a Chomskyan — the above was meant to be a joke. Sorry if that didn’t come through clearly.
I got it, SC.
Incidentally, I just tried the Grammar Interactive Quiz
http://www.ccc.commnet.edu/grammar/challenge.htm
available through one of the links on the Language Log post. Back to class for me, I fell two short of the 15 needed for a pass!
My favorite bit of theoretical nonsense from a Chomskyan.
background: graduate syntax class taught by a hardcore Chomskyan (very nice person and truly gifted lecturer despite the unfortunate affiliation)
Anyway, the lecturer was discussing the idea that every sentence has to have a recoverable subject. That is, according to his Noamness, at some level of forumalation, every sentence has to have a subject which can be added to the surface level representation so that using a simple example from Thai
deep representation
Hpom m’ai r^uu khrab
i not know {polite particle}
-hpom (deletion of hpom)
surface representation
m’ai r^uu khrab.
In other words, even when hpom is dropped in the surface representation, it’s still there in some way and can be recovered easily enough.
I countered with the famous(?) example of the Polish -no/-to construction
Rozpoczęto nową budowę.
begun new construction
Construction has begun.
Nie słuchano mnie.
not listened me
I wasn’t listened to./No one listened to me.
Although this is called the impersonal passive in Polish and although it’s most often translated as an agentless passive in English, it’s not really a passive at all as the object remains in whichever object case is required by the verb. Furthermore, no noun or pronoun in the nominative case can appear in the sentence, not even a dummy subject like ‘to’ is allowed. I mentioned all this to the lecturer and was told (paraphrasing, this was a long time ago)
“Well, counter-examples don’t disprove the theory that this is universal.”
That’s classic! Thanks for the much-needed laugh.