We have seen an example of a grammatical syllabus, and noted that structures are presented in lists. At this point, it is worth asking how and why these lists are organized, and why, for example, the present tense is presented before the past tense. The justification for listing was noted nearly 100 years ago:

To ' know' a grammatical form is to be able to use it freely; it will be undesirable therefore to introduce the pupil to too many forms at once. The language must unroll itself before him gradually. It is for this reason that we begin with the present tense, and pass from that to the perfect and the future. (Kittson, 1918: 104)

Indeed, Mackey also points out that grading in language education has existed for at least 500 years, and notes that Comenius was the first to claim

that systematic gradation reduced the difficulties of language learning by distributing the extensive material of a language into steps arranged in specially prepared text in which everything progressed, not by leaps and bounds, but gradually. (Mackey, 1965: 204-05)

Willis (2003) also accepts a role for the gradual accumulation of language items, at least at lower levels of proficiency.

Typically, grammar-based approaches involve the use of contrived or scripted language to illustrate the target structure. Many believe that such contrivance facilitates comprehension, and would worry about the use of unmediated authentic texts (by unmediated, I mean texts which are engaged with by the learner without guidance/interpretation from the teacher):

The scripted text is unreal English, which is unlikely to be reproduced in actual contexts of use but is easier to comprehend, and more real pedagogically; the unscripted text is real English, but more difficult to comprehend and to produce, and therefore likely to be considered less real pedagogically. (Carter, 1998: 47)

A key principle, alongside the question of grading and sequencing, is simplification, the idea that sentence-level examples can be contrived to provide simple and easily understandable instances of target structures. Simonsen (1987, cited in Nation, 2001) has suggested that, at least until recently, this has been a policy used by publishers in producing graded readers, but that the simplification process has been dependent on writers' intuitions rather than, say, frequency of occurrence of lexical and grammatical patterns. Simonsen also notes that 'some publishers make sure that grammatical and discourse signals are as explicit as possible, particularly with regard to pronoun reference, direct and indirect speech, and conditional sentences' (Nation, 2001: 171).

The traditional language syllabus, consisting of lists of discrete and simplified grammatical rules accompanied by contrived linguistic illustrations of these rules, is 'still used in the vast majority of classrooms the world over' (Long and Robinson, 1998: 15), and the reasons are pretty much as we have already set out:

Language - any language - seen from 'outside', can seem to be a gigantic shapeless mass, presenting an insuperable challenge for the learner […]. By tidying up language and organising it into neat categories (sometimes called discrete items), grammarians make language digestible. (Thornbury, 1999: 16-17)

This implies that 'a language teaching course can be based on a sequence of linguistic categories, and […] that these categories can be exemplified in simple sentences for intensive practice' (Howatt, 1984: 141). What we are really talking about in one form or other is Structuralism/audiolingualism where the sentence reigned supreme. The sentence-level language description system of Structuralism remains with us in contemporary foreign language textbooks.

But how is the standard language syllabus to be sequenced and graded? For Thornbury (1999: 9-10) grading the syllabus should include:

• Complexity: an item is complex if it has a number of elements: the more elements, the more complex it is.
• Learnability: the learnability of an item was traditionally measured by its complexity: the more simple, the more learnable.
• Teachability: the fact that it is easy to demonstrate the meaning of the present continuous ('I am walking', 'she is writing' etc) has meant that it is often included early in beginners' syllabuses, despite the fact that it has a relatively low frequency of occurrence (Thornbury, 1999).

For myself, I would include the following:

• Frequency: the importance of frequency as a constraint in textbook design has been noted by Willis (1990), and with the use of Corpus Linguistics techniques we are now in a position to include structures which are frequent (there is little point, for example, in spending much time teaching infrequent structures like the 3^rd conditional)
• Utility: structures should be useful in communication acts (the 3^rd conditional is not very productive or useful, and learners are unlikely to need to produce very often sentences like, 'If I had had money, I would have bought a car'.
• Complexity: the number of items/elements in a structure ('he'll be arriving late' has more structural elements and is therefore more complex than 'he's arriving late'); the number of operations in a construction may also be important, both within a language and across languages. The formation of interrogatives in Portuguese, for example, involves zero operations, ie intonation provides the interrogative impact, and there are no structural modifications:
  
  

  João	comeu	o chocolate.
  John	eat-past	the chocolate.
  'John ate the chocolate.'

  
  

  O João	comeu	o chocolate?
  John	eat-past	the chocolate.
  'Did John eat the chocolate?'

  
  Japanese on the other hand, uses one operation to form interrogatives, the insertion of the question particle, 'ka':
  
  

  John wa	chokoreto wo	tabemashita.
  John-subject	chocolate-object	eat-past
  'John ate the chocolate.'

  
  

  John wa	chokoreto wo	tabemashita ka.
  John-subject	chocolate-object	eat-past-question particle
  'Did John eat the chocolate?'

  
  English is the most complex of the three languages, using two operations, the breaking down of the verb into stem and auxiliary, and subject-auxiliary inversion:
  John ate the chocolate.
  Did John eat the chocolate?
• Truth: language patterns should be true to life, modelled on or extracted from authentic language; again, the use of corpora allows for this with relative ease.

Finally, Swan (1994: 45-52) lists the following criteria for pedagogic rules:

1. Truth: rules should be true.
2. Demarcation: a pedagogic rule should show clearly what are the limits on the use of a given form.
3. Clarity: rules should be clear.
4. Simplicity: a pedagogic rule should be simple.
5. Conceptual parsimony: an explanation must make use of the conceptual framework available to the learner.
6. Relevance: a rule should answer the question (and only the question) that the students' English is 'asking'.

The typical grammatical syllabus consists, then, of isolated lists of grammar items, all taught in a predetermined sequence devised in accordance with the criteria we have set out above. Often, these grammar items are illustrated by the use of specially contrived sentences or texts.