Take the first point. Someone may hold out an opened pack of cigarettes and ask.“Smoke?”or a person may answer the question“Do you smoke?”by means of the one-word“Yes！”or the question“which one of these boys was seen to smoke?”by“Johnnie！”and so on. Countless other example are possible. In every instance，we are clearly dealing with eilipsis. The single word utterances are only interpretable by virtue of the listeners ability to supplement the omitted part of the sentence. The first instance is interpreted as the sentence,“Do you smoke?”,the second as“I smoke”(or, “I do smoke；hence“Yes, I do smoke.”)；and the third as“Johnnie has been seen to smoke.”There may ,in some instance, be ambiguity because not enough context is given to enable the listener to place the single word into the intended sentence. But generally it is correct to say that the meaning of word is uninterpretable in social commerce, unless we have enough clues with to construct a sentence for that word.
The second point is factually correct：utterances heard in colloquial English(or any language, for that matter)do not conform to what we know to be correct grammar. We must make here a distinction. There are indeed utterances that are totally“ungrammatical,”but they are also uninterpretable－we do not know what the speaker was trying to say. On the other hand, much more often we do know what the speaker wanted to say even though his utterances are clearly ungrammatical. This may be because he omitted part of the sentence or because a sentence is begun as if it were to end in one way but is actually concluded by using the second half of a different type of construction.(Several variations of this are possible.) Our capacity to understand such semi-sentences can only be due to a facility to supplement the omitted part of incomplete sentences. Thus, the interpretation of semi-sentences is not simpler than the understanding of grammatical sentence but actually requires a special ability：to supplement the missing parts of a partially concealed pattern(analogous to pattern-completion in visual perception).If a sentence is under certain circumstances﹝Osgood(1957) cites the example,“Garlic I taste！”﹞, this is not necessarily a sign that syntax may be abandoned at will rather of the existence of correspondence that do not ordinarily enter into the writing of normative grammars. The rule of correspondence in this case relates the form of the“Garlic”sentence to the form of such sentences as“I taste garlic！”The example cited is not necessarily an instance of agrammatism but merely that of an admissible rule. That the types of such rules are limited (or that the rules have a psychological reality) is seen in the fact that the words in this sentence cannot be permutated in all possible ways.
In the light of this discussion, how do we explain the onset of language development where it is a universal finding that children begin with one-word utterance? Does this mean that the observations on adult language are false? Or that they are irrelevant? I do not believe that either us the case. To the contrary, if we assume that child’s first single word utterances are, in fact, very primitive, undifferentiated forms of sentences, and that these utterances, and that these utterances actually incorporate the germs of grammar a number of phenomena may explained.
There is a period at which an infant may have a repertoire of up to 50words including such items as daddy, here, milk, up, baby, etc. He will utter any one of these words in isolation and they may mean：Daddy, come here；Daddy went by-by；no milk；more milk, please； etc. But even though the child’s memory is sufficient to know all of the 50 words. and even though he hears such phrases as here is your milk, shall daddy take you by-by, etc., he will neither join together any two words he knows nor can he be induced to do so upon request. This cannot be explained by assuming that he makes himself better understood this understood this way, that the reference of the words(that is, the association with the object)is still too narrow and fixed；or that he has no need for putting words together；or that he cannot vocalize for that long a period of time；or that thus due to poorly developed general memory. All of theses assumptions are refutable by observations. Nor would any of theses assumptions make it clear why the child suddenly and spontaneously does begin to join words into two-element phrases.
有一個時期嬰兒有高達50個字，包含一些項目，例如：daddy, here, milk, up, baby等等，他將孤立的發出這些字的任一個來表達意思，例如：Daddy, come here；Daddy went by-by；no milk；more milk, please等等，但就算小孩的記憶力可以記下50個單字，甚至他聽的一些片語例如：here is your milk, shall daddy take you by-by等等，他既不會把他知道的單字串聯在一起也不會試圖詢問，這不可以經由他讓自己變的更好的方法來解釋，這些相關的字(包含物體)實在太狹窄及固定，或他不需要把這些字串聯在一起，或他不可以發出這麼長的時間，或由於記憶發展的很緩慢，這些假設都可經由觀察被駁倒，這些假設都不可清楚的解釋小孩爲什麼可以突然自動的進入雙語時期。
The assumption that the early single word utterances are primitive syntactic units－in a sense primitive sentences－find support in the following considerations. Semantically, and in terms communication, the single words seem to function in the same way that sentences come to function later on：they cover a complete proposition：for instance, they may stand for a statement such as, Daddy is coming down the street. phonologically they may be operated upon by a given rule. Much the way a whole string of symbols is operated upon later on；for example, one of variety of intonation patterns the utterance such as declarative, interrogative, or hortative pitch-contours. It is reasonable to assume that the formal processes that regulate the perception and production of sounds are essentially the same as those that enter into syntax and that the one-word stage is simply a transitional stage during which the rules are extended from the interaction of articulatory movements to the interaction of larger language units, namely morphemes words, and that the eventual acquisition and mastery of grammar has its origin right at the beginning of language development；otherwise we would have to assume that some day the child “discovers”grammar and makes an effort to learn this phenomenon, which seems farfetched.
關於早期單字發聲的假設者要是以句法為單位-一個早期句子的知覺，從下列的原因有支持的證據，在語意上，會話時單詞的功能和句子之後的功能相似：她們包含了完整的主題：例如：他們可能會產生一個像Daddy is coming down the street.的句子，在語音上，他們會產生一個規則，產生一連串的記號使以後可以管理，例如：會利用一些圖示來表多樣的語調、疑問詞、或音調的輪廓，這就是爲什麼我們確信正常語音感知的調節、產生過程和構成語法是一樣重要的，且單語時期是延長構音運動到比較大的語言單位的過渡時期，專有名詞，最後才習得文法規則，且是從初的語言演變來的，反則我們將確信小孩某天就突然「發現」文法及習得語音，但這太牽強了。
(3) Theoretical Considerations
(a) Understanding-Speaking. It is easier to construct a theory that explains why adults understand sentences the way they do, than a theory that explains why or how a given seguence of words is produced by a specific person at a specific time. This is not say that understanding language is based on a separate mechanism from producing language. Both are based on the same apparatus of principles. But if we test an individual. There must be certain motor capacities, memory, motivation, a specific train of thoughts, given social conditions and factor.
It is easier to study general capacities for behavior than the specific forms that behavior will take at any one time, and it is easier to predict the capacity for understanding than the capacity for speaking, because there are fewer factors affecting the former than the latter. A similar point, but with further refinement, is made by Chomsky, Appendix A, under the headings of competence and performance.
研究行為普遍的能力是比研究特定行為在某時刻的形成簡單、且預測理解的能力是比預測說話的能力簡單，這是因為影響前面的因素地影響後面的因素少，相同地，從更近一步的研究中，由Chomsky, Appendix A提出在下方的標題-形成能力。
The distinction made here is relevant to many types of behavioral studies. Suppose we wanted to make a psychobiological study of chess playing. For instance, we wanted to know ,“What are the mental characteristics necessary for this game?”or, “Can a chimpanzee learn to play it?”The empirical questions that would be asked in this research are：can a given subject learn the various moves? Can he develop a strategy? Does he see the implications of his adversary’s moves? And so on. We to know whether he can comprehend the game. If we had nothing but a catalogue of his moves without a report of what his opponent was doing(that is, how he understood his opponent’s game), we should have an imperfect idea of his competence as a chess player.
That the understanding of language is more relevant to an estimation of language capacities may also be seen from the following：we can learn to understand a language without ability to speak it. This is true of primary language acquisition, as well as the acquisition, as well as the acquisition of a second language. In these cases, the underlying principles of the language are acquired, but the development of the skills for production are lagging.
There is an interesting situation here which would be a paradox unless we were willing to make assumptions on the nature of the learner: as investigators of the nature of language, it is preferable to concentrate on understanding; the objects that are to be understood are sentences; the sentences that are actually heard are frequently “degraded specimens ’’from a grammatical point of view-they are semisentences at best; on the other hand, the understanding of semisentences is apparently more difficult proper sentences. We first seem to learn the rules and principles underlying grammatically correct sentences, and only by virtue of having acquired these can we begin to understand semisentences. (This becomes particularly evident in the acquisition of a second language by a good lecturer or presented in print, long before we can understand a conversation which is heavily loaded with semisentences ) . The paradox is this: if the child’s task is to abstract principles that generate correct sentences, but is presented indiscriminately with semi- and proper-sentences, how can the correct principles be established, and why or how does his understanding of sentences become fairly explainable in terms of a grammatical theory? The assumption that we have to be willing to make here concerns the cognitive machinery that we must suppose to be developing in the child.
這裡有一個有趣且矛盾的情勢，除非我們願意在學習者的本性上假定︰作為自然語言的調查者，比較能專心於理解，而目標是理解句子；實際上聽到的句子經常是「被降級的樣本」，從文法的觀點上來看他們是不完全句子（semisentences）； 另一方面，理解這些不完全句子顯然是比理解那些句法完整的句子要來的困難。我們一開始似乎會先學習規則以及文法基礎來修正句子，利用習得的優點讓我們開始理解不完全句子。 (這顯然成為一個演講者或是書寫者習得第二語言的方法，很早以前我們就用來理解含有大量不完全句子的口語)。矛盾的是：假如孩子對於收集正確句子的原則是很抽象難懂的，不過隨意呈現不完全句子以及完全句子，他是如何建立正確的原則，還有為什麼或者如何在文法理論中運用他的句子理解能力而能夠完全解釋？這項假設讓我們願意將這樣的關係把認識文學橋段與我們推想應該是發展於兒童連結在一起。
If the most promising source-material for a theory on syntactic mechanisms is understanding, what data should we use to construct a theory on the development of syntax in children? His actual utterances may, in certain cases at least, be irrelevant to his development of syntactic mechanisms(for instance, in children with severe psychiatric disease who may not choose to speak, or who prefer to make animal noises, or in children whose noises cannot be understood). By and large it is true that young children can understand more than they can say.＊
Children between 18 and 36 months seem to have a tendency to run constantly through their repertoire of capacities. This is also reflected in their verbal behavior in that during this period the gap between their understanding and speaking capacity normally remains fairly constant and predictable. This may be tested by asking them on the one hand to execute certain verbal commands or to point to pictures that are being described to them in more or less complex sentences, and on the other hand to require them to repeat accurately sentences that are given them. Since a sentence contains so much detail we cannot repeat it correctly upon a single presentation unless we can apply grammatical principles to it by means of which the mass of information can at once be recorded and thus processed in much simplified from Mehler and Miller, (1964). The utterances of a child who is just beginning to speak (normally not much later then 30 months) may thus reflect the stages that his development of language capacities, particularly understanding have traversed, even though one may actually have taken place some 2 months before the other. By about 30 months, however, production soon becomes as unreliable an indicator of language capacities as is the case in the adult. Unfortunately, no studies have yet been published that have undertaken systematic research on the development of grammatical understanding of the child at this age and older. Even the best studies have relied too heavily on production.
18~36個月大的小孩似乎習慣經常使用理解能力範圍的內容。他們的理解和說話能力常態來說是可預期的，這也是這段期間表現在言辭上的行為。這或許可以用問的方式來做測驗，一方面實行固定的言辭上的命令，或者是指出圖片並且使用或多或少的完整句子來描述他們，另一方面要求他們準確地重複句Mehler 和 Miller(1964)表示，因為一個句子包含很多細目，所以我們在無法正確地一次的發表中重複它，除非我們可以藉由對照資訊的語法規則立刻紀錄以及使過程簡單化。一個剛開始說話的小孩(常態來說不會晚於30個月大)，他的語調或許反應了語言發展能力的階段，尤其是理解的經過，甚至事實上取代兩個月前就進行其他了。然而，大約30 月，語言能力的指標生產不久成為不可靠，變得像成年人那樣。不幸地是，沒有研究發表承擔有系統對的研究語法理解孩子的發展在和年紀更大的這年齡。即使在生產上倚賴最好的研究已經太嚴重。
(b)How Mature Speakers Understand Sentences. Some insight into this problem is provided by asking ourselves why a sentence such as
They are boring students has two meanings. Here the explanation is quite simple; we may choose to link the word boring to the word are as in Fig.
Each of the circles may be characterized by an abstract name such as shown in Fig. 7.4. Instead of writing the names on cylinders or circles we might set up a list of definitions such as shown in Table 7.2. Similar information as that shown in Fig. 7.4 as a cylinder-diagram and in Table 7.2 as list of definitions is represented by Chomsky in the form of a tree diagram which he calls phrase-marker. The grammatical principles illustrated in Figs. 7.3 and 7.4 may be diagrammed by phrase-markers as in Fig. 7.5.
每個圈都可能如圖7.4中所示以摘要詞性為特點。 不是寫那些名字在汽缸上或者環繞如表格7.2中所示我們可能建立這些定義的目錄。在圖7.4顯示一張汽缸圖解為相似的訊息 並且在表格7.2列出Chomsky定義的目錄稱為片語製造者的一張樹狀圖。在圖7.3以及圖7.4裡說明的語法原則。在圖7.5 片語製造者可能用圖表示。
 But Roger Brown (personal communication), in his extensive investigations of the first steps toward language acquisition, has found that this is not always and necessarily true. For example, there were instances in his sample in which plural inflections were used productively at time when experiments on the child’s semantic progress indicated that he did not yet know what this particular suffix signals. Similar observations were described by Fraser et al. (1963).