On the December 17 edition of this website, I had an article published which detailed a sentence – “Who will be easy for us to get his mother to talk to?” - I’ll call it Sentence A - which a number of Chomskyan scholars were analyzing, seemingly unaware that the sentence was ungrammatical. I saw this as a failure of Chomskyan linguistics, in that a first rule of linguistic analysis is to determine the grammatical status of a structure under consideration.
There were quite a few viewer comments to that posting, some of which were favorable, and some of which questioned my motives in attacking Chomsky.
One of those who disagreed with me was an individual who identified him/herself only as “a Chomskyan linguist.” I will refer to this quite knowledgeable person as “Chom Ling”. The point that Chom Ling wanted to make was that my understanding of the issues was incorrect, and so, using the comments section of the FrontPageMag.com website, we engaged in a debate that lasted for well over a week. I thought that Chom Ling fought a fair fight, raising objections and challenges which were reasonable and appropriate, and acknowledging that I had scored a few points. I will note that at no point did Chom Ling ever introduce into the debate any concepts or terms derived from Chomskyan theory. That seemed strange; after all – not to put too fine a point on it – if the tools which are part of one’s espoused theoretical framework are not useful in order to help one argue or advance one’s own argument…oh, never mind.
Anyway, after this had gone on for a week or so, Chom Ling suggested a cessation of hostilities – a Christmas truce - on the grounds that the grammaticality of the sentence was ultimately a matter of personal judgment: I viewed the sentence as ungrammatical, and Chom Ling and others viewed it as grammatical, and so there was no way to bridge the gap. Or, in the words of Chom Ling: “you…maintain that it's ungrammatical. That's much harder for me to prove false; it's a statement about your own reaction to the sentence, and that's a domain where you know best.”
Spoken like a true Chomskyite.
I observed the Christmas truce and then the next day I sent Chom Ling a somewhat more formal proof that showed a number of grammatical impossibilities in the sentence. (The proof, slightly revised, is reproduced below.) I claim, on the basis of that proof, that the grammatical impossibilities render the sentence demonstrably ungrammatical. The Chomskyan scholars had indeed been studying a sentence which they did not realize to be ungrammatical.
I never heard back from Chom Ling, and so I have to assume that no flaws in my proof were found. No concession was offered but that was not surprising: from my experience, Chomskyans don’t concede; they simply fall back and regroup. Presumably Chom Ling is still regrouping.
The idea that there was no point in continuing the battle because ultimately it was a matter of individual judgment is a notion that is at the heart of Chomskyan linguistic theory. The scientific methodology upon which Chomskyan linguistics is based is the idea that whatever a native speaker of a language “feels” to be right is grammatical. In other words, grammaticality is determined by personal intuition. This doctrine has become so thoroughly ingrained in the dominant schools of modern linguistics that no one questions it, and yet it has got to be one of the goofiest ideas ever recorded in the long annals of science. This idea runs counter to all the hard sciences, which require that evidence be paramount. What would the field of physics, or chemistry, or mathematics look like if personal intuition trumped considerations of evidence?
No doubt this conviction was derived from the perfectly apposite notion that all languages are equally capable of expressing complexity and are equally capable of infinite nuance. But that is a far cry from saying that whatever a speaker says – as long as it seems right to the speaker – is grammatical. People produce ungrammatical sentences all the time, either because they do not know the rules of their language, they choose to ignore them, or they simply can’t perceive any logical contradiction within the sentence structure. (If you don’t believe me, just ask the president.)
Anyway, Sentence A was looked at by any number of scholars – including Chom Ling - and nobody suspected any problem with it in terms of grammaticality. And yet once the sentence was fully analyzed it was shown to be grammatically impossible – on three different counts. It was ungrammatical because it violated three logical rules of English grammar: 1) A verb can have only one meaning in a sentence. 2) A verb can be either active or passive, but not both. 3) If the object of a verb is expressed grammatically, it must be the same as the logical object. In Sentence A, all of those rules were violated.
And so Sentence A was ungrammatical, not because I ‘felt’ it to be ungrammatical, but because it violated various rules of logic in the structure of English. It is ungrammatical no matter how many people say otherwise.
This Chomskyan notion of personal intuition as the basis for grammaticality, though long-accepted, is nevertheless highly problematic. For example, what do we say about a child who says, “Me want cookie”? Surely the child must feel comfortable with this structure, and must feel that it will convey the right idea. And we can certainly extract the meaning out of the utterance. Therefore, using Chomskyan precepts, the child must be speaking grammatical English. (I leave aside the point that, if grammar were innate, as Chomsky claims, the child would intuitively know that the case for the pronoun subject ought to be nominative.)
Consider as well a native Russian living in America who has been speaking English for years and is fully conversant. The Russian may have heard Americans use definite articles a myriad of times and yet the Russian says, “I have key to apartment.” and he would feel this to be grammatical, since the Russian language does not use articles, as English does. (The Russian feels no more need to say “the key” or “the apartment” than we do to say “the France”, although if we were speaking French we would have to add the article to be correct.) Since we can easily understand what the Russian means, and since he feels that he is expressing all the necessary ideas without the use of articles, then, again by the Chomskyan methodology, the Russian must be speaking grammatical English.
And yet neither the child nor the Russian is speaking grammatical English, because English requires that subjects be in the nominative and that specificity be indicated by articles or other means.
Or consider the NPR announcer who recently said, “The Ukrainian government will allow whomever wants to vote…” No doubt “whomever” sounded acceptable to the NPR announcer, but he ought to have checked before inflicting such an excruciating construction on millions of listeners. (“want” has to have a subject. Thus: allow… those who want/anyone who wants/whoever wants…to vote.)
Or consider the following well-known grammatical phenomenon. There are many people who would never say, “Me went to the races.” but who would nevertheless say, “Me and Larry went to the races.” According to Chomskyan precepts, both are grammatical because they both would seem to be intuitively correct to some speakers. Now if both sentences are grammatical, then we should be able to extract a rule which governs this situation. The rule would have to be something like this: ‘A single subject pronoun must be in the nominative case, but a pair of subject pronouns may be in the accusative case.’ The problem with this rule is that there is no logical reason that the number of subjects should determine the case of the subjects, and so we are forced to conclude that the scientific methodology of Chomskyan theory leads us to the formulation of grammatical rules which are unsupported by logic. And so this house of cards must fall.
I should further comment that this notion - that any human utterance is grammatical as long as the speaker thinks that it is - breeds a kind of anti-intellectualism, whereby there is no particular compulsion to rigorously discern whether a structure is grammatical. We need only discern whether the speaker thinks it is acceptable, and then an attempt is made to create a grammatical framework which incorporates it.
Thus, in Chomskyan linguistics, sentences which violate the rules of the language are lumped in with sentences which don’t, in an attempt to create some overarching theory which explains all human utterances. Sentence A in the Linguistic Inquiry article is a good example of that: the various scholars twisted themselves into knots trying to reconcile the properties of the ungrammatical Sentence A with various other sentences, some of which were grammatical and some of which were not. The catalogue of arcane terms employed in the article in order to try to reconcile the grammatical with the ungrammatical is extensive: weakest crossover configurations…bound variable anaphora…asymmetric linking…licensing conditions…etc.; to the contrary, the only concepts I needed in order to show that Sentence A was ungrammatical were: active versus passive voice, the notion of a logical object versus a grammatical object, and the idea of semantic content.
And so this anti-intellectualism, which requires that natural human utterances not be looked at too closely in order to determine whether a given structure comports with the rules of the language, has over time evolved into a kind of pseudo-intellectualism, whereby a grand theory is under perpetual construction and revision – a theory which will explain the properties of all natural human utterances. I would suggest that any such endeavor is a waste of time, as it is impossible to create a systematic description of structures, some of which are themselves systematic, and some of which are not.
If not even the experts can tell without careful examination whether a given structure is grammatical or not, then this suggests that it is time to abandon the Chomskyan methodology which allows personal intuition to be the determiner of grammaticality. The medical model is a good one to follow: the patient describes the symptoms, but it is the doctor who provides the diagnosis. The first order of business ought to be to determine whether or not a given structure is grammatical. This is done not by taking a vote, or by asking someone’s opinion, but by determining whether the elements of the structure are in logical conflict or not. It is not asking too much to require that a verb be either active or passive, but not both. It is not too much to ask that the case of the subject be determined by something other than the number of subjects in the sentence.
The Chomskyan goal of creating a theory which incorporates all natural human utterances ought to be abandoned as well. Once the ungrammatical structures are culled, then the grammatical structures can be studied with a view towards developing a coherent theory. Of course, one might find that once the grammatical structures are separated from the ungrammatical, theories of grammar will seem to be superfluous. A grammar which is complete in itself requires no theory to explain it. And perhaps that is really the point: if the Chomskyans had spent the past forty years trying to understand the difference between a grammatical sentence and an ungrammatical one, they might have had a great deal more to show for their effort.
Proof: That the following sentence is ungrammatical: “Who will be easy for us to get his mother to talk to?”
1: The verb phrase “to get S to” means: to cause someone to. The verb phrase “to get X” means: to obtain X or to understand X.
2: The adjectival phrase “easy to V” where V is an active verb, is interpreted as a passive: e.g. (1) The plane is easy to see. means (2) The plane is easy to be seen (= The plane is easily seen.) Also, (3) An avalanche is easy to cause = (4) An avalanche is easily caused. Thus, the verb takes an active form but has a passive meaning. This characteristic we can designate: F(a)M(p).
In (3) and (4) “avalanche” is both the grammatical subject of these sentences and the logical object of “cause”.
In (5) It is easy to cause. and (6) It is easy to be caused., we can allow “it” to take the place of “avalanche”. The subject “it” refers to semantic content: avalanche.
4: We can also render the meaning of (5) as (7) It is easy to cause an avalanche. In (7) the logical object “avalanche” is no longer referred to in the subject of the sentence, but as the grammatical object of “cause.” Thus, the following differences between (5) and (7) are in effect:
(a) the subject “it” no longer contains semantic content;
(b) the verb is now active in both form and meaning: F(a)M(a); and,
(c) the grammatical object and the logical object refer to the same thing.
5. The sentence “Bob is easy to get.” is grammatical in and of itself. It follows the structure: easy to V, where V = to get X. Thus, since the V following “easy to” must be interpreted as a passive, the meaning of the verb is “is gotten”.
Since there is no grammatical object expressed, the logical object is "Bob”, which is the same as the subject. Thus, the meaning of the grammatical sentence “Bob is easy to get” means “Bob is easily gotten.” If we choose to express the grammatical object of the verb, then we could say, “It is easy to get Bob.” where the subject loses its semantic content.
6. If we take the sentence “Bob will be easy to get.” and add “his mother to talk to” then the verb “to get” has an object which is grammatically expressed: his mother. Note, however, that if the object of “to get” is grammatically expressed, then it must be the same as the logical object (See 4(c) above.) But “Bob” and “his mother” do not refer to the same person, so we have a logical impossibility. The only way to allow the logical object and the grammatical object to refer to the same thing is to have the subject become semantically empty, as in (7). But the subject cannot become semantically “emptied” of Bob because that semantic content has no place to go since Bob is not the same person as his mother.
If the semantic content “Bob” has no place to be transferred, then it must stay in place within the subject. Since the logical object is the same as the subject, we have a passive meaning, and we have determined that in such a case with “easy to” the verb takes an active form but has a passive meaning, and is designated: F(a)M(p). And yet the grammatically expressed object “his mother” which follows “to get” requires that the verb be of the F(a)M(a) type. (See 4(c) above.) Since semantic content in the subject requires F(a)M(p) and a grammatically expressed object following “to get” requires F(a)M(a), we have a second grammatical impossibility.
Also, we noted that the meaning of the grammatical sentence “Bob is easy to get” means “Bob is easily gotten.”
But if “his mother” follows “to get”, then the verb means “to cause.” Thus, the verb has two meanings within the same sentence, and that is a third grammatical impossibility.
Thus, the sentence is ungrammatical on three counts.
7. If “Bob will be easy for us to get his mother to talk to” is ungrammatical, then the sentence “Who will be easy for us to get his mother to talk to?” is also ungrammatical, since the pronoun “who” stands for “Bob.”