An Introduction to Applied Linguistics

bucket, hell bent for leather), in that they have associated discourse functions. They 
range from completely fixed, as in by the way, which functions to shift a topic in 
discourse, to relatively fixed frames with limited slots for fillers, as in a___ago, used 
to express time relationships (for example, a day ago, a long time ago), to frames 

---

26 An Introduction to Applied Linguistics
with slots allowing considerable variation, as in I’m (really) (very) sorry to hear that 
X (where X may be an entire clause, such as, you flunked the test, you lost your job, 
etc.), used to express sympathy.
The descriptive part of the problem is that these phrasal units, which are 
pervasive in language, cannot be adequately accounted for by models consisting of 
abstract rules of sentence syntax, supported by a lexicon of single word items that 
are inserted into abstract categories such as NP, VP, PP, etc. There is considerable 
evidence that the mind stores and processes lexical phrases as individual wholes, 
including evidence from first language acquisition studies indicating that they 
are learned first as unanalysed chunks and, only later, analysed as to particular 
grammatical patterns (Peters, 1983).
At present, there is growing interest in investigating the implications of 
formulaic language for descriptions of grammar, in particular, implications for 
how we view the components of syntax and lexicon and for how the components 
interact with each other and with discourse level concerns (DeCarrico, 1998). A 
closer look at the limitations of various grammatical models may help us to re-
examine previous assumptions and to look for new directions in resolving issues 
and problems in the description of grammar. As this essential work on arriving at 
more comprehensive descriptive grammars continues, applied linguists must also 
get on with the tasks of explaining the learning, and improving the teaching, of 
grammar.
Learning Grammar
Over the history of applied linguistics, different theories of learning have been 
proposed to account for how grammar is learned. During the middle of the 
previous century, for instance, grammar learning was thought to take place 
through a process of verbal ‘habit formation’. Habits were established through 
stimulus-response conditioning, which led to the ‘overlearning’ of the grammatical 
patterns of a language. In order to help students overcome the habits of their 
native language and inculcate those of the target language, teachers conducted 
pattern practice drills of various types: repetition, transformation, question and 
answer, etc. Teachers introduced little new vocabulary until the grammatical 
patterns were firmly established. Language use was also tightly controlled in order 
to prevent students making errors that could lead to the formation of bad habits 
that would later prove difficult to eradicate.
With the rise of generative grammar and its view of language as a system of rules, 
grammar learning was seen to take place through a process of ‘rule formation’, 
which itself was brought about when students formulated, tested and revised 
hypotheses about grammatical structures in the target language. Thus, students 
were seen to play a much more active role in the classroom than they had earlier. 
Consistent with this perspective, students’ errors were not to be feared, but rather 
welcomed as evidence that students were attempting to test their hypotheses 
and receive feedback, with which they could then revise their hypotheses. In the 
classroom, students were given written grammar exercises so they could induce 
the grammatical rules that would allow them to generate and understand novel 
sentences.
With the shift toward a more communicative approach to language teaching, 
views of grammar learning changed once again. Some held that grammar 
learning took place implicitly and most effectively when students’ attention was 

---

27
Grammar
not on grammar at all. In other words, they said that grammar was best learned 
subconsciously when students were engaged in understanding the meaning of 
the language to which they were introduced (Krashen and Terrell, 1983). Those 
that adhered to a Chomskyan universal grammar (UG) perspective felt that target 
language input alone or input with negative evidence (that is, evidence that a