“Real exceptions” do not exist
“What is ‘grammar’?” series: Part 3
As we discussed previously, grammar, to linguists, is systematic. Every language is bound by a set of systematic rules known collectively as its grammar.
Once again, “systematic” means that if a certain rule or restriction exists in one situation with a particular set of conditions in the language, the same rule or restriction would apply in another situation with the same set of conditions as well.
Because the rules that make up the grammar of a language are always systematic, there are never any “real exceptions” to them—as in, it would never be the case that something is allowed in one sentence but not another if the relevant conditions are exactly the same. For example:
X I think ^ knows*.
We encountered the sentence above in Part 2, when we first discussed how the “systematic” rules of grammar restrict all sentences with the same conditions.
This sentence has an embedded clause within an outer clause. We can see the structure of the embedded clause more clearly by indicating clause boundaries with square brackets [ ].
X [ I think [ ^ knows* ]. ]
We saw previously that the embedded clause here does not have a subject. This sentence is incorrect because the lack of a subject in the embedded clause violates a systematic rule in English grammar--the one that forbids the subject of a finite clause to be dropped. Here, the finite verb in the embedded clause is “knows,” but this finite verb does not have a subject.
Now, once we know that the sentence above is incorrect because its embedded clause does not have a subject, we would know that any sentence that has an embedded clause with no subject is also incorrect. This is because the rule that “a finite clause cannot drop its subject” is a systematic rule in English grammar and binds all sentences with the same conditions:
X [ They believe [ ^ will succeed* ]. ]
X [ Everyone tells me [ ^ is too late ]. ]
Because we know the systematic rule that a finite clause cannot drop its subject in English, and we know that “I think ^ knows*” is incorrect, we would know that these two sentences above are incorrect as well. Like “I think ^ knows*”, these two sentences also have an embedded clause with no subject—“^ will succeed*” and “^ is too late*.”
Equipped with the knowledge of this simple systematic rule in English grammar, we would be able to identify the problem with these sentences by ourselves: as long as one of these sentences is incorrect, the other two must also be incorrect, because all three sentences share the same conditions in terms of having a subject-less embedded clause. We do not need a teacher to point out to us that these sentences are problematic each time we write them.
Because grammar is systematic like this, there would never be an “exception”—a “real exception”—in the sense that one sentence with a subject-less embedded clause is incorrect while another one is correct.
However, sentences that we encounter in real life are often not so clean-cut. We often encounter examples that seem to be “exceptions” to certain established rules and restrictions in the systematic grammar of a language.
In these situations, we might find that two sentences seem to share the same conditions, but one of them is bound by a systematic rule while the other one is not.
Understanding that grammar is fundamentally systematic means that, when we see that two sentences are not bound by the same rule, we know that their conditions must somehow be different, even if they seem identical on the surface.
What linguists often have to do is to figure out this difference between seemingly identical situations and sentences. This is often extremely complex, and different linguists might even disagree--but, the bottom line is, there must be a difference between two situations if the same rule does not apply to both of them.
In other words, “real exceptions” never exist in grammar.
Next, we will use an analogy to further illustrate how we should think about the “seeming exceptions” to grammar that we encounter.
文法並沒有「真正的例外」
「grammar」（文法）是甚麼系列：第 3篇
早前我們已談論過，對語言學家而言，「grammar」（文法）是系統性的。每種語言都受一套系統性的規則規範，我們統稱這些規範為「grammar」。
再次強調，「系統性」是指在該語言中，如果某規則會限制或規範擁有某個特定條件的情況，則這規則也定必會限制或規範另一個擁有相同條件的情況。
由於組成文法的規則都是有系統性的，所以「真正的例外」其實並不存在，一定不會出現某規則規範一個句子，卻不規範另一個相關條件完全相同的句子的情況。例如:
X I think ^ knows*.
本系列的第二節中，當我們第一次討論文法的「系統性」規則如何約束所有條件相同的句子時，也使用了上述例句。
這句有一個被外層子句包裹著的嵌入式子句 （embedded clause）。我們在博客、YouTube和網上課程中，都使用了[ ] 括號將子句分界，讓大家能看清嵌入式子句的結構。
X [ I think [ ^ knows* ]. ]
這嵌入式子句中如果沒有主語，這句子就必定是有結構錯誤的，因為它違反了英文「grammar」中的一個「系統性」規則 － 限定子句中的主語不能被省略。在這嵌入式子句中的限定動詞是「knows」，但並沒有配合它的主語。
從這個嵌入式子句中沒有主語這一點，我們明白這句子有結構錯誤的；那麼我們也應該知道，所有包含有不帶主語的嵌入式子句的句子都會是錯誤的。
因為「限定子句中的主語不能被省略」是英文文法中的一條系統性規則，它可以約束所有相關條件相同的句子：
X [ They believe [ ^ will succeed* ]. ]
X [ Everyone tells me [ ^ is too late ]. ]
由於我們知道「限定子句中的主語不能被省略」是英文文法中的一條系統性規則，以及知道「I think ^ knows*」是錯誤的，那麼我們也必然知道上述兩個例句都是不正確的。正如「I think ^ knows*」一樣，以上兩句也包含了嵌入式子句，而嵌入式子句中並沒有加入主語去配合限定動詞「^ will succeed*」和「^ is too late*」。
清楚了英文文法中這個簡單的系統性規則，我們便不需要每次都靠老師指出問題，而是懂得自行發現這些句子的錯處 － 只要以上其中一個句子是不正確的，另外兩個也一定不正確。因為這三個句子的共通點，是它們都包含不帶主語的嵌入式子句。
「grammar」的系統性使它永遠不會出現「真正的例外」 － 也說是說，一個不帶主語的嵌入式子如果在某一句中是錯誤的，在另一句中都一樣會是錯誤的。
然而，我們在現實生活中看見的句子往往不會都如此清晰簡潔。我們總會發現一些句子，乍看之下像是系統性文法中某些特定規則的「例外」。
有時我們可能會發現兩個似乎擁有相同條件的句子，其中一個受著某系統性規則的規範，但另一個卻沒有。
理解「系統性」其實是「grammar」的根本，我們便會明白這些句子中的條件必定有某程度上的不同，無論它們表面看來有多相似。
語言學家經常要從那些表面相似的情況和句子中找出差異。這是十分複雜的工作，語言學家之間甚至也未必會互相同意。但共同的底線是，如果同一規則不適用於兩種情況，兩種情況之間必定存在差異。
換句話說，就是「grammar」系統中從不存在「真正的例外」。
下一節，我將以一個比喻來進一步說明，我們應該如何思考和理解這些文法中「看似例外」的情況。