The meaning of a sentence is determined form its deep structure by means of the rules of semantic component of a grammar. The phonetic form of a sentence is determined from its surface structure by means of the rules of phonological component. Thus, grammar may now be defined as a system of rules for relating pronunciation of a sentence to its meaning. The syntax, or the base, is at the core of the system and generates the infinite number of structures underlying the grammatical sentences of a language. These structures are, in turn, given a semantic and phonetic interpretation by means of the semantic and phonological component of the grammar. The phonological and semantic component are purely interpretative since they do not generate any sentences on their own but rather assign the phonetic and semantic form to the sentences generated by the syntax.

The base comprises of a set of categorial rules as well as of a lexicon. These two combined serve a function similar to that of phrase structure rules of the earlier system of TGG, with minor exceptions. The lexicon includes all the vocabulary in the language and associates with each the syntactic, semantic, and phonological information essential for the correct operation of the rules. This information is termed ‘features’. For instance, the syntactic features for ‘boy’ are: [+ Noun], [+ Count], [+ Common], [+ Animate], and [+ Human]. The categorial rules have two basic functions, namely “they define the system of grammatical relations, and they determine the ordering of elements in deep structures” (Chomsky 1965:123). The categorial rules generate phrase markers with a number of slots to be completed with lexical items. There is a set of features associated with each slot, which features determine the kind of item that can be filled in the slot. For instance, the noun ‘boy’ could be filled in a slot specified as requiring an animate noun, a common noun, a human noun, or a countable noun, however, the very noun would be excluded from slots specified as requiring an abstract noun (e.g. ‘wisdom’) or an uncountable noun (e.g. ‘sugar’). With the aid of syntactic information (features) assigned to the lexicon, the categorial rules permit sentences such as ‘The boy laughed”, while rejecting and defining as ungrammatical non-sentences such as “The boy elapsed”.

To conclude, Standard Theory modifies the original Transformational-Generative Grammar Theory in that it introduces to the system of grammar an additional semantic component represented by a deep and surface structure in a sentence. Moreover, Standard Theory makes a sharp distinction between competence and performance of an ideal speaker-hearer of a language.

Within the framework of Extended Standard Theory, a view that deep structure exists as a distinct level of syntactic structure at which lexical insertion occurs, and which serves as the input to the transformational component is maintained. Nevertheless, the rules of semantic component operate not on deep structures exclusively, as it was the case in Standard Theory, but also on surface structures of sentences.

According to Chomsky “linguistic theory is concerned primarily with an ideal speaker-listener, in a completely homogeneous speech-community, who knows its language perfectly and is unaffected by such grammatically irrelevant conditions as memory limitations, distractions, shifts of attention and interest, and errors (random or characteristic) in applying his knowledge of the language in actual performance” (Chomsky 1965:3). The study of actual linguistic performance should account for a number of factors, the underlying competence of the speaker-hearer in particular. Thus, competence can be defined as the speaker-hearer’s knowledge of his language which allows him to construct and understand indefinitely many grammatical sentences. Performance, on the other hand, can be defined as the actual use of language in concrete situations and is concerned with acceptability rather than grammaticality. It should also be borne in mind that errors occurring in linguistic performance are irrelevant to the study of linguistic competence since performance does not directly mirror competence. The goal and challenge at the same time for a linguist and for a child learning a language is to figure the underlying system of rules of the speaker-hearer on the basis of his performance. Thus, a grammar of a language may be said to aim at a syntactic description of the ideal speaker-hearer’s intrinsic competence, and if this grammar is explicit enough, it may be termed a generative grammar. Hence, generative grammar is there to describe the knowledge of the language of a speaker-hearer, which knowledge he utilizes in actual language use.

Generative grammar is a system of rules which generate an indefinitely large number of structures. These rules may be analyzed within three components of generative grammar, namely syntactic component, phonological component, and semantic component. The syntactic component comprises of rules which generate sentences and assign to them two structural analyses, a deep structure and a surface structure. Deep structure is represented by the underlying phrase marker, which is assigned by rules of the base (an equivalent of phrase structure rules of the early model of TGG), whereas surface structure is represented by the final derived phrase marker, which is assigned by the transformational rules. Figure 23 below presents how the four sets of rules interrelate.



Fig. 3 Simple phrase structure grammar rules (Chomsky 1957:26)
Each of the rules (i) – (vi) in the above diagram has the form of X → Y, where X is a single element, Y is a string consisting of one or more elements, and the arrow suggests that the element to its left be replaced (rewritten) with the string of elements to its right. Considering what follows:

Fig. 4 Derivation of the sentence “the man hit the ball” (Chomsky 1957:27)
one shall assume the above diagram as a derivation of the sentence “the man hit the ball”. The numbers to the right of each derivation line stand for the rule of the grammar presented in Fig. 3, which rule was applied to create that line from the preceding one. Thus, line two in Fig. 4 is formed from the first line by substituting Sentence with NP + VP with reference to the rule (i) in Fig. 3, line three in Fig. 4 is formed from the second line by substituting NP with T + N with reference to the rule (ii) in Fig. 3, line four in Fig. 4 is formed from the third line by substituting VP with Verb + NP with reference to the rule (iii) in Fig. 3, and so on. The derivation of the sentence “the man hit the ball” can be demonstrated more clearly by means of a tree diagram (also known as phrase-marker) as presented below:

Fig. 5 Tree diagram (Chomsky 1957:27)
It must be, though, stated that the tree diagram reveals less information than the derivation in Fig. 4 due to the fact that it fails to show the order, in which the rules were applied in Fig. 4. Only by means of the derivation in Fig. 4 can one construct the tree diagram in Fig. 5, but not reversely, for it is possible to construct a derivation that reduces to the tree diagram in Fig. 5 with a different order of application of the rules. With the use of the tree diagram in Fig. 5 one can notice what is crucial in Fig. 4 in order to determine the phrase structure (constituent analysis) of the derived sentence “the man hit the ball”. Provided that the sequence of words of this sentence can be tracked back to a single point of origin in the tree diagram in Fig. 5, then a sequence of words of this sentence is a constituent of type Z, and its points of origin are labeled Z. Therefore, “hit the ball” can be tracked back to VP in Fig. 5 and as such it is a VP of the derived sentence. However, “man hit” cannot be tracked back to a single point of origin in Fig. 5, and as such it is not a constituent at all.
If two derivations reduce to the same diagram of the form presented in Fig. 5, they are equivalent. If a grammar constructs nonequivalent derivations of the same sentence, it is a case of ‘constructional homonymity’ (structural ambiguity) and, provided that the grammar is correct, this sentence of the language should be ambiguous, thus analyzable in more than one way. For instance, the sentence “flying planes can be dangerous” may be classified as ambiguous since ‘planes’ might be either the object of ‘flying’ or the subject of ‘can’ (Chomsky 1957).
It is now essential to generalize on the rules of the simple phrase structure grammar presented in the diagram in Fig. 3. It is important to be able to limit application of a rule to a certain context. Hence, T may be rewritten a on condition that the subsequent noun is singular (but not plural), analogously, Verb can be rewritten “hits” on condition that the preceding noun is man (but not men). In order to limit the rewriting of X as Y to the context Z – W, one ought to formulate the following rule: Z + X + W → Z + Y + W. For instance, considering singular and plural verbs, instead of creating an additional rule Verb → hits in the diagram of Fig. 3, we ought to have NP sing + Verb → NP sing + hits demonstrating that Verb is rewritten hits only in the context of NP sing . Similarly, the rule (ii) of the diagram in Fig. 3 will have to be reformulated so as to account for NP sing and NP pl. Therefore, the rule (ii) of the grammar of the type presented in Fig. 3 could be replaced by a rule such as:

Fig. 6 NP sing / NP pl rule (Chomsky 1957:29)
where Ø stands for the morpheme singular for nouns and plural for verbs (e.g. “boy”, “come”), and S stands for the morpheme singular for verbs and plural for nouns (e.g. “comes”, “boys”). However, there is one feature of the grammar of Fig. 3, which must be retained, namely only a single element can be rewritten in any single rule (Chomsky 1957).
One point of generalization that may now be made is that each grammar based on constituent analysis is defined by a finite set Σ of initial strings as well as the finite set F of rules of the form X → Y. Although X does not have to be a single symbol, only a single symbol of X may be rewritten in formulation of Y. Analyzing the grammar of Fig. 3, it is possible to state that the single symbol Sentence was the only member of the set Σ of the initial strings, and F comprised of the rules (i) – (vi), however, one could modify Σ so as it would account for Declarative Sentence or Interrogative Sentence as additional symbols. Within the framework of the grammar [ Σ, F ], one can now define derivation as “a finite sequence of strings, beginning with an initial string of Σ, and with each string in the sequence being derived from the preceding string by application of one of the instruction formulas of F” (Chomsky 1957:29). It is now essential to introduce the term ‘terminated derivation’, that is a derivation whose last string cannot be further rewritten by the rules F. Consequently, the last string of a terminated derivation is its ‘terminal string’, and the set of terminal strings of the grammar [Σ, F] is a ‘terminal language’. The phrase structure of each sentence, which is a terminal string, can be reconstructed by means of associated diagrams, which may also define the grammatical relations in the terminal languages.
Providing that it is possible to generate all the grammatical sequences of morphemes of a language by means of the grammar [Σ, F], the very grammar should hence be completed by a phonemic structure of these morphemes in order for it to produce the grammatical phoneme sequences of the language. This might be achieved by creating a set of rules of the form X → Y, for instance, for English, in a form similar to the kind presented:

Fig. 7 Morphophonemic rules (Chomsky 1957:32)
and incorporating it into the phrase structure derivations. One should, however, bear in mind that it is no longer required that only a single symbol be rewritten in each rule of the set of morphophonemic rules illustrated above. Given the above, the application of Fig. 7 to the phrase structure derivations, will produce a unified process of generating phoneme sequences from the initial string Sentence, thus making it possible to generate the grammatical phoneme sequences of the language.
The phrase structure grammar based on the immediate constituent analysis has been so far proved to be more adequate than a model based on the finite state processes. However, attempting to consider sentences beyond the scope of the simplest type of phrase structure grammar, one finds the process extremely difficult and complicated. In order to examine the limitations of phrase structure grammar, some examples depicting improvements that ought to be made over the grammar of the form [ Σ, F ] should be presented.
It is well known that, among many other processes, sentences may be created by means of conjunction. Given the two sentences of the form Z + X + W as well as Z + Y + W, where X and Y are constituents of these sentences, it is possible to create a new sentence of the form Z – X + and + Y – W. Considering the following:

Fig. 8 Sample sentences / conjunction (Chomsky 1957:35)
one should see that sentence (21) can be formed from sentences (20a-b). However, when X and Y are not constituents (or are constituents but of different kinds in a sense that each of them has a single point of origin in a tree diagram, though this origin is labeled differently), we cannot use the above form to create sentences of the type considered. For instance, we cannot form (25) from (24a-b).

Fig. 9 Sample sentences / conjunction (Chomsky 1957:36)
The description of conjunction may be simplified by setting up constituents so that the rule presented below will apply.

Fig. 10 Conjunction rule (Chomsky 1957:36)
Although it is much simpler to account for conjunction by means of the above rule rather than without such a rule, we are, at this point, prone to encounter difficulty arising from the fact that due to certain limitations of a grammar [Σ, F] of phrase structure, and the like, it is impossible to incorporate the rule of Fig. 10 is such grammars. The basic feature of the rule of Fig. 10 is that one ought to know both the form of S1 and S2 as well as their constituent structure prior to the application of the rule to sentences S1 and S2, form which a new sentence S3 is to be formed. In this respect, it is clear that the very rule fails to apply to phrase structure grammar since it does not allow such double reference.
Another example can illustrate improvements that ought to be made over the grammar of the form [Σ, F]. Fig. 3 accounts only for one way of analyzing the element Verb as hit. It is obvious that the verb may take a number of other forms, for instance, some of the forms of the verb take are: takes, has + taken, will + take, has + been + taken, is + being + taken. Equally obvious is the fact that analysis of the aforementioned auxiliary verbs is an important process in the development of the English grammar. However, once attempted to include these phrases into a [Σ, F] grammar, one discovers that the description appears to be rather complex. In order for the grammar of the type presented in Fig. 3 to account for the occurrence of auxiliary verbs in, say, declarative sentences, it should include the following rules:

Fig. 11 Auxiliary verb phrase rule (Chomsky 1957:39)
The interpretation of the rule (28iii) is as follows: “we must choose the element C, and we may choose zero or more of the parenthesized elements in the given order. In (29i) we may develop C into any of three morphemes, observing the contextual restrictions given” (Chomsky 1957:39). Not reiterating the initial steps, the following derivation of the type of Fig. 4 may be constructed so that it illustrates the application of the rules of Fig. 11.

Fig. 12 Derivation of the sentence “the man has been reading the book” (Chomsky 1957:39-40)
Every auxiliary verb phrase can be generated in a similar way, perhaps with a necessity for adding additional rules that would ensure grammaticality of the sequences to be generated.
However, in order to apply the rule (29i) in the derivation of Fig. 11, one had to make reference to previously conducted derivations to be able to state the constituent analysis of the NP sing the + man, and as already seen in case of the rule of Fig. 10, such double reference collides with the requirements of the grammar of the type [ Σ, F ].
Yet another example illustrating improvements that ought to be made over the grammar of the form [Σ, F] concerns active and passive sentences. In forming passive sentences one makes use of the element be + en of the rule (28iii), though there are certain restrictions that have to be accounted for before the element be + en can be selected, namely the Verb that follows needs to be transitive, Verb cannot be followed by a noun phrase, if Verb is transitive and is followed by the prepositional phrase by + NP, then it is crucial to apply be + en before the string V + by + NP thus inverting the order of the adjacent noun phrases. Hence, in order to extend the grammar of the Fig. 3, it would be necessary to pose many restrictions on the element Verb so that this grammar would produce only grammatical sentences. Those restrictions, on both the elements Verb and be + en, might be excluded provided that one reintroduces passives from the phrase structure grammar by means of the following rule:

Fig. 13 Passive rule (Chomsky 1957:43)
Although the above rule significantly simplifies the grammar, it is well beyond the scope of the grammar [Σ, F] in that it is necessary to make reference to the constituent structure of the string to which it is to apply and it imposes an inversion on this string.

Chomsky (1957) claims that English is not a finite state language since it seizes to construct a finite state grammar that would account for all and only grammatical sentences. As he further states, finite state grammar fails to explain certain processes of sentence formation in English. It is possible that some no-adjacent words of a sentence will hold certain amount of dependency and such a dependency may be further separated by a phrase or clause containing string of non-adjacent interdependent words. This, in turn, results in a sentence with ‘mirror image properties’. Considering a sentence with a following structure a + b + c … x + y + z, there is dependency between the outermost constituents (a and z), between the next outermost (b and y) and so forth. Any language that contains an indefinitely large number of sentences with ‘mirror image properties’ similar to the above example is outside the scope of the finite state grammar. Chomsky did not, however, reject finite state grammar theory as being inapplicable to the study of language, but he rather claimed that ‘left to right’ and ‘word by word’ generation of sentences would not account for some of the constructions in English. Thus, he provided more powerful tool for the study of language, phrase structure grammar, which comprised of all what finite state grammar was capable of doing – and much more.
References:
Chomsky, N. 1957. Syntactic Structures. The Hague: Mouton.
Lyons, John. 1970. Chomsky. Fontana.

When rules of grammar of the target language are talked about, this approach is referred to as explicit grammar teaching. It requires the students to learn rules given by the teacher, afterwards practicing them in exercises. Since explicit teaching of grammar involves talking about rules prior to showing examples and doing exercises, lessons focusing on depicting the rules might be boring or oppressive for students. Moreover, they frequently fail to be communicative. This approach to teaching might be effective in the case of learners having logical-mathematical intelligence. In addition, Purpura and Bachman (2004) state that explicit knowledge is assimilated at a slow pace.

The second way of teaching grammar is implicit, which means that learners are involved in this process in a passive way in that they are exposed to the language, work with it and grammar rules are not mentioned. The underlying assumption is that the learners will spot the regularities and assimilate items of grammar simultaneously being aware of their meaning. Due to the fact that this approach excludes thorough explanation, it is thought to be suitable for child who has short attention span (Purpura and Bachman 2004:42).

It is at early childhood that these patterns of a native language gain significance which, in later development, are transferred to second language acquisition in an attempt to identify them in similar situations. Where the native language and the learned language differ in some aspects the transfer of patterns might not be very productive. Should such a situation occur, help of a teacher is of use. Evidence provided by methodologists supports that theory, in addition indicating that students learn a foreign language focusing on the whole of the language as well as on its parts (Lynne 2001:15).

One other advantage of learning a foreign language in young age is children’s evident willingness to try to talk in the target language which is useful when it comes to acquiring pronunciation (Lynne 2001:1).

Child not only can learn directly from instruction, but also indirectly through interacting with others from what they can interact with. That is why the majority of tasks for young learners ought to contain movement and stimulate various senses. Using different activities by child in acquiring foreign language is a crucial factor which makes his learning fun (Scott and Ytreberg 1990:5).

He presents a new idea named LASS which stands for Language Acquisition Support System. Brumfit, Moon and Tongue (1995:3).

“Burner is one of the most notable contemporary exponents of the view that children’s language and learning development takes place through the process of social interaction. […] Burner proposed that for language development there needs to be a child component, incorporating an innate propensity for active social interaction and language learning, together with an adult support and help component. The international partner provides a structure of framework which Burner referred to as scaffolding”.

This idea is possible thanks to the concepts which Bruner named as: formats and routines which by combination of a well – known situation with anticipation of something new enable scaffolding to occur (Lynne 2001: 9). Language that child produces during the routine is known, however a new situation which may not be easy to predict, gives opportunity for child to produce his own chunks of language. This space perfectly corresponds with Vygotsky’s ZPD, and what is more, Bruner adds that routines and their tuning with “the new” provides chances for cognitive development. Recognizable situations provide chances for linguistic improvement, additionally allowing child to develop languages. In other words, routines serve as opportunities for expanding knowledge (Lynne 2001:11).