by Ramesh Krishnamurthy

Unit 1 - From Lexis to Grammar

1.0 Introduction

As language is the subject matter of our teaching, it is vital that we have an in-depth understanding of Grammar and Lexis, as they are key aspects of language, whichever curriculum or syllabus we are following, and whichever model of language we are applying.

As teachers, it is also important that we adopt relevant strategies to achieve our teaching goals. As Dave Willis’s opening sentence says: ‘Whenever we do anything in the classroom we are acting on our beliefs about language and language learning’ (Willis 2003:1).

So, for example, if you present your students with the item cat, and do not make any mention of the forms cats and cat’s, you (like most dictionaries) have implicitly made the judgment that cats and cat’s are to be considered as grammatical features rather than vocabulary items, and that these forms are to be learned from the grammatical rules about the formation of plurals in English, or the ways that possession is indicated in English. On the other hand, if you present learners with concordances from a corpus containing the forms cat, cats and cat’s, you are explicitly showing them the different forms that are associated with the item cat and the set of meanings which cat can evoke.

It is such underlying judgments and beliefs which will come under the spotlight in this module, and hopefully by examining them and questioning them, you will emerge with a better idea of why you teach the things you do, and in the way you do, and you should also find some interesting suggestions for new things to teach and new ways to teach them.

In this Unit, we will look briefly at the historical attitudes towards Lexis and Grammar, and their treatment in linguistics and language teaching, including an overview of the contributions of lexicography and the recent advent of electronic corpora. We will discuss some of the technical terminology used in the study of lexis, and consider to what extent individual words constitute an appropriate unit for the study of language. We will look at some of the problems in defining lexical units, and consider the process of vocabulary acquisition from a pedagogical viewpoint. We will then discuss the notion of grammar, and explore some of the different kinds of grammar.

1.1 The Status of Grammar and Lexis: a bit of historical background

1.1.1 Lexis is currently prominent in language teaching

In Unit 10 of FND, titled Lexicogrammar, Julian Edge said: ‘lexical perspectives on language, language learning and language teaching have made up the growth area in our professional field for the last fifteen years or so’. So, to some extent, this module reflects the current thinking in the field, in allotting 6 units to Lexis and 3 to Grammar.

Julian also referred to ‘relatively recent developments in the use of computers with insights into grammar and language teaching which have arisen from a renewed focus on words’ (Edge 2004). This partly explains both why we are starting with lexis before moving on to grammar, and why we will be covering aspects of corpus analysis in this module.

But why has lexis become prominent so recently, and why was it not so before?

1.1.2 Historically, linguistics focused on grammar, and neglected lexis

Linguistics traditionally centred on grammar. For example, Robins says that linguists regard phonology and grammar (often subdivided into morphology and syntax) as the core concerns of linguistics, with some argument about the status of semantics and phonetics (although he actually chooses to include both; Robins 1964:18-19).

The explanation Robins gives for the exclusion of lexis is that ‘The categories of phonetics, phonology and grammar are general; the components of the lexicon of a language are particular’, so the lexicon ‘requires particular and different statements for each item’, and is therefore described by the linguist Bloomfield as ‘an appendix of grammar and the list of basic irregularities’ (ibid:63).

So lexis was for a long time regarded as the province of dictionary-makers, not linguists.

And teaching followed the path of linguists, not lexicographers. As Roe (1998:Unit 1:1) points out, ‘We may in the past have suffered from over-subservience in the face of the “great” linguists’, with the result that language teaching, language courses and coursebooks were designed on the basis of grammar.

If grammar was the focus in teaching, the main unit of grammar was the sentence. I certainly remember learning French and Latin at school as a series of sentences, which had to be translated, or transformed in some way (singular to plural, active to passive, etc). Lessons were a sequence of verb conjugations (e.g. -er verbs, -ir verbs,
-re verbs, etc in French and 1st, 2nd, 3rd, 4th conjugations, etc in Latin) and noun declensions (Latin), the imperative mood, subjunctive mood, and so on. Grammar was also the focus of the coursebooks, with short texts specially devised to focus on the topic of the lesson. Lexis, on the other hand, was much more do-it-yourself. There were brief glossaries in some of the coursebooks, but we were also issued with blank vocabulary books, in which we were supposed to write the words and phrases we came across.

The teaching of English as L1 was subject to the same structural analysis, based on the Latin model. We learned the parts of speech, ‘box-analysis’ (clause functions such as Subject and Object), and the subjunctive mood (‘If I were…’). We were taught that cat’s was genitive case, me was the accusative form of the pronoun, and so on. The prescriptivist rules that we were taught said that we should not use split infinitives, put a preposition at the end of a sentence, or begin a sentence with and or but.

The problem was that I was constantly coming across examples while reading, listening to the radio, watching TV, or in conversations (i.e. examples of ‘use’), that did not fit in with the ‘system’ I was being taught. The system was being imposed, and I could no longer believe that all the examples I was encountering were just ‘exceptions that proved the rule’. There were just too many of them. So I gradually realized that the problem was that the system was inadequate, or simply wrong.
I would rather trust the examples of ‘use’ and try to find a system which explained them more comprehensively and convincingly.

1.1.3 Firth, Halliday and Sinclair: lexis becomes a part of linguistics

J.R. Firth (1890-1960) placed semantics at the core of linguistics: ‘Indeed, the main aim of descriptive linguistics is to make statements of meaning’ (Firth 1957a:190), and said that the meaning of ‘language events’ could be dealt with at different levels: from social context through syntax and vocabulary to phonology and even phonetics, or in the opposite order (ibid:192).

As Halliday said of Firth:
‘At a time when few linguists, other than lexicographers themselves, devoted much attention to the study of lexis, and outlines of linguistics often contained little reference to dictionaries or other methods in lexicology, J.R. Firth repeatedly stressed the importance of lexical studies in descriptive linguistics. He did not accept the equation of `lexical' with `semantic', and he showed that it was both possible and useful to make formal statements about lexical items and their relations.’ (Halliday 1966:148).
Lexis was being recognised as an autonomous level of language.

That 1966 paper by Halliday was called, significantly, Lexis as a Linguistic Level.
And in the same collection of papers was one by Sinclair, entitled Beginning the Study of Lexis. It is interesting to note that Sinclair, writing in 1970 (Sinclair et al 1970/2004:3), still feels that Halliday has not yet sufficiently accommodated lexis,
and refers to ‘Halliday (1961), where lexis was assigned the role of picking up the scraps from the tables of syntax’.

By the way, Firth was also important for incorporating the social dimension of language into linguistics, no doubt sharing the interests of his contemporary anthropological colleagues, especially Malinowski, with whom he developed the concept of ‘context of situation’ (Robins 1964:27).

Halliday built on this social aspect in his Systemic Functional Grammar, in which the ‘functional’ part refers to how the language is used (cf. my comments at the end of section 1.1.2 above), in a way rather similar to the ‘context of situation’. You will find this topic discussed in much more detail in Units 7-9 of this module, which focus on Hallidayan grammar.

The term ‘lexico-grammar’ is now often used in recognition of the fact that lexis and grammar are not separate and discrete, but form a continuum.

1.1.4 Lexis was made prominent by Language Teachers, EFL Dictionaries, and Corpora

Language teachers also had a hand in promoting the importance of lexis from the 1930s onwards. As Cowie (2000) tells us, Palmer was interested in vocabulary control (or limitation) because he thought he could make language learning easier by ‘pinpointing those relatively few words which carried the main weight of everyday communication’. Palmer was joined in his lexical studies later by Hornby.
Meanwhile, West wanted to produce simplified readers for schoolchildren. These three teachers produced word-lists (e.g. Palmer and Hornby’s Thousand Word English, and West’s General Service List) and the first monolingual learner’s dictionaries (West’s New Method English Dictionary, and Hornby’s Oxford Advanced Learner’s Dictionary).

Owen (1993:171) refers to Carter and McCarthy (1988:42), who cite Wilkins’ book Linguistics and Language Teaching (1972:111): ‘Without grammar very little can be conveyed, without vocabulary nothing can be conveyed’.

The advent of corpora (which we shall discuss in much more detail in Unit 2), with their natural lexical orientation, has pushed lexis to the forefront of lexicography, language studies and language teaching. See Sinclair (1987, 1991), Willis (1990) and Lewis (1993) for more details.

1.2 Why do we call the subject ‘Lexis’, and not ‘Vocabulary’, or just ‘Words’?

Why do we use three different terms to refer to the same thing?

I assume that you have all read the Foundation (FND) module, before starting on Grammar and Lexis (GLE).

In FND, the terms word/words, vocabulary and lexis were used many times.
Let us look more closely at how these terms were used in FND and in what contexts.

In order to do this, I have designed a set of three short tasks which
should help you to understand better:

a) why ‘lexis’ is preferred as the name of the academic subject
b) the relationship between words, context, and meaning (which we will discuss in more detail in Unit 2),
c) how modern lexicographers compile dictionaries from corpus examples (more on this in Unit 2, too).

As you read through the examples in each table for the first time, concentrate on exploring the meaning(s) of the term in bold. Then do the task in the box that follows each set of examples.

1.2.1 Word/words

Examples from FND Notes/Comments
1. Your response should not exceed one thousand words.
2. First of all, that word tangible which I decided to use in the previous paragraph.
3. Lynch is using the word noticing as a technical term from the field of second language acquisition.
4. I use the word in this sense below when I refer to ‘fellow professionals’ in the section on Voice and relationships.
5. Previous friends took against some of the ‘posh’ words that I started to use.
6. Let us imagine for a moment a classroom in which the teacher says to the students, ‘Put the chairs in rows.’ Given the presence of some kind of recording, the saying of these words can be established pretty unproblematically as objective fact.
7. In other words, how do we choose to specify syllabus content?
8. Memorise the above questions roughly and work them into the exchange in your own words.

TASK 1.1: Word/words
a) Make your own notes/comments on each of these examples.
b) Does each example reflect a different use (or meaning) of word, or can you put the examples into ‘use/meaning’ groups?
c) Write a rough definition for each of these uses/meanings.
d) Which elements in the context did you think were significant?
e) Does it matter that I have given examples of both word and words? Would it make the task easier if we considered the examples of word and words separately?
f) Can you think of three other uses/meanings of word, not represented here?
g) Write example sentences for these uses.
h) Now look up word in a learner’s dictionary.
i) How many uses are listed?
j) Can you think of any missing uses?
k) Did any of the information given in the dictionary surprise you?

1.2.2 Vocabulary

Examples from FND Notes/Comments
1. We all do this anyway to some extent, adjusting our choice of vocabulary and style of speech to the situation we are in.
2. Using pictures to pre-teach vocabulary in reading classes with young learners in Thailand…
3. A similar point can be made about the lists available under the headings of Functions, Notions and Topics, Grammar, Pronunciation and Vocabulary.
4. …deciding how to go about learning some new vocabulary.
5. … helpful resource materials for the learners’ own use, such as grammar-reference sections, vocabulary lists, tape transcripts, answer keys…
6. …produce a questionnaire to investigate vocabulary learning.

TASK 1.2: Vocabulary
a) Make your own notes/comments on each of these examples.
b) Does each example reflect a different use/meaning, or can you put the examples into ‘use/meaning’ groups?
c) Write a definition of vocabulary for each use/meaning you found.
d) Substitute word or words in the examples above.
e) What effect do you perceive in each case?
f) Try to account for this effect.
g) Now look up vocabulary in a learner’s dictionary.
h) Does the entry match your opinions?

1.2.3 Lexis

Examples from FND Notes/Comments
1. Show also any clear signals of the functions that you see, such as sub-headings or key lexis.
2. …appreciate the role of lexis in signalling discourse organisation.
3. …strictly grammatical words such as ‘the’ and ‘if,’ and fully lexical words such as ‘cat’ and ‘beautiful.’
4. Furthermore, their actual meaning in any given instance has to find ‘lexical realisation’ in the text…
5. It is now often argued that a focus on lexical development (words and phrases) is far more suitable for beginners and Young Learners.
6. Computer-generated concordances can provide the basis for a lexical approach to the study of grammar.
7. …predict lexical meaning from intuition and check against data.
8. …discuss the idea of a lexical basis for syllabus, materials and methods.

TASK 1.3: Lexis
a) Make your own notes/comments on each of these examples.
b) I have included examples of both lexis and lexical. Does this confuse or help you to understand the concept they both relate to? Is it different from including both word and words in the first set of examples? Why?
c) Does each example reflect a different use/meaning, or can you put the examples into ‘use/meaning’ groups?
d) Write a definition of lexis/lexical for each use/meaning you found.
e) Can you think of any other uses of lexis/lexical?
f) How does the use of lexis/lexical differ from word/words and vocabulary?
g) Can you now explain the difference in use between the three terms?
h) Look at the entries for lexis and lexical in a learner’s dictionary.
i) Do the entries match your opinions?

The tasks you have just completed (1.1 to 1.3) are the type of tasks that are often set as training exercises for lexicographers. Are you now tempted to change careers?

1.2.4 The final word on words, vocabulary and lexis

I hope you will have realised by now that the main reason why we don’t use word/words as a technical term is because of its ambiguity. Although I have only shown you examples from one single text, written by the same author, and belonging to a very restricted genre (academic, pedagogic writing), word/words is still used with many meanings. If we had looked at a much wider range of text-types, we would have found many more meanings, as the entry in your learner’s dictionary no doubt indicated. For example the Cobuild dictionary (2001) gives 38 uses, including the use as a verb (to word something in a particular way) and various phrases.

Vocabulary does not seem to have as many different uses. Cobuild (2001) lists only 3:
1. your vocabulary (the words you know)
2. all the words in a language
3. the vocabulary of a subject (words typically used to discuss a subject)

But FND usage suggests that vocabulary is often used in contexts involving words like learning or teaching, and therefore may be the preferred pedagogic term.

Lexis is given the label ‘technical’ in Cobuild (2001), but lexical is not. Do you think this is appropriate? I must say that I found it odd, so I checked the 450-million-word version of the Bank of English corpus (part of Collins Word Web, copyright HarperCollins), which I have access to via Birmingham University, and discovered that lexis occurred only 36 times (mostly in linguistics lectures), whereas lexical occurred 135 times and in a wider range of source texts. So perhaps the lexicographer was correct after all. Lexis has only one use, as a technical term in linguistics. Lexical has only one use, but can also be found in non-linguistics texts.

This is why linguists prefer to use the term lexis. It is unambiguous, and is rarely used outside linguistics.

1.3 Are ‘words’ an appropriate unit for the study of language?

Task 1.4: Re-read pages 1-10 of Singleton (2000) carefully at this point, before moving on to the next section of the module.

1.3.1 ‘Words’ are (of course) used as a unit of language in linguistics

Singleton (2000:1-3) raises a point that is relevant to my earlier discussion of the historical neglect of lexis by linguistics: although lexis was not considered to be a legitimate category of linguistics, of course words formed the underlying basis for the discussion of many of the categories that were included.

For example, phonology uses minimal pairs of words (Singleton gives the examples of pin/tin, top/tot, and gape/gate for English) to test for phonemic distinctions in a language (in this case /p/ and /t/).
Also, both aspects of grammar, morphology and syntax, depend on the existence of the word – morphology for word-formation, syntax for word order in clauses or sentences (although ‘word order’ here may actually refer to the order of groups of words rather than individual words).

And semantics (which Robins included, but many other linguists do not accept as a linguistic category) often starts with (and certainly usually deals with) word meanings, as well as higher-level statements such as ‘propositional’ meanings.

1.3.2 Some other technical terms for ‘word’

Singleton also mentions another set of technical terms which we should pay attention to: token, type, lexeme and word-form.

Singleton (2000:5-10) looks at various ways in which ‘word’ can be defined, and introduces the terms token and type. These terms are often used in corpus linguistics, so it is useful to know what they mean.

A token is any word in any text.

In the previous sentence, there are 8 tokens: ‘A(1)-token(2)-is(3)-any(4)-word(5)-in(6)-any(7)-text(8)’. Singleton describes tokens as ‘actual occurrences of any items that might qualify (as a word)’. Some linguists refer to these as ‘running words’.

Every different word in a text is a type.

The sentence defining token above, which itself contained 8 tokens, contains only 7 types: ‘A(1)-token(2)-is(3)-any(4)-word(5)-in(6)-any(4)-text(7)’, because there are two occurrences of the word ‘any’. Singleton describes types as ‘items with different identities’, and gives the example of the phrase ‘going(1)-going(1)-gone(2)’ which has only 2 types, but 3 tokens. As Singleton points out, we have begun to ‘group or classify’ items. And in the case of type, the grouping is because of identity of spelling.

A ‘family’ of related words, such as go, goes, going, went, gone, is sometimes called a lexeme. This is similar to the normal dictionary concept of a word. A lexeme consists of several word forms, and the headword in a dictionary is often referred to as the ‘citation form’.

Some linguists call the lexeme a lemma and its forms lemma-forms.

However, the main problem with using ‘words’ as a unit for language study is that it is not a precise unit. Singleton indicates some of these imprecisions.

1.4 Defining a ‘word’ precisely in technical linguistic terms is difficult

Task 1.5: Read more on this topic in Singleton (2000) Chapter 6.

1.4.1 Defining ‘word’ by Spelling

Singleton refers to this as the orthographic approach: ‘a word is a sequence of letters bounded on either side by a blank space’. This is a purely formal definition (meaning that it depends solely on the written form of a word). Singleton mentions some problems with this approach.

a) this way of defining a ‘word’ works for English, but not for some other languages

It works reasonably for written English (and some other languages, depending on the script they use, e.g. Roman, Cyrillic) but not for others, for example Chinese (in which each character represents a concept) and Japanese (which uses Chinese characters as well as two other sets of characters).

Task 1.6: As we are really only dealing with English in this module, do we need to worry about other languages?

If it works well for English, I do not think we need to worry about other languages.
We are not ultimately linguists wanting to describe all languages. We are English teachers wanting to teach English well.

Task 1.7: Can you think of some of the ways in which this definition does not work, even for written English? Take some time to reflect on this question before you move on.
(Some of the problems with this definition are discussed below)

This definition is in fact the initial basis for corpus analysis, because it is nice and simple, and computers can deal with it. However, many decisions and adjustments have to be made first, because we don’t want to lose valuable linguistic information.

Some of the problems with defining ‘word’ by spelling are as follows.

Some problems with Tokens:

(i) punctuation: we don’t put a space between a word and a following punctuation symbol, so we need to eliminate punctuation. But this would cause problems with some abbreviated forms: should B.B.C. become 3 tokens (B B C) or one (BBC)?

(ii) apostrophes are another problem: we would not want cat’s to become cats, but do we want to treat cat’s as 2 tokens (cat and s) or one (in which case, we would have to preserve the apostrophe in some way). Another problem is that apostrophe and the single-close-quote symbol is often the same character (e.g. The Jones’ cat and We watched the TV programme ‘Meet The Jones’). How can we disambiguate them?

(iii) hyphenation: nowadays, the rules for hyphenation are usually created and implemented by computer programmes (spellcheckers, computer typesetting of books, newspapers, books, etc). Linguistically, there are many inconsistencies.

Some problems with Types:

(i) capitalization: do we really want to distinguish between sentence-initial forms and non-sentence-initial forms of the same word (e.g. Apple and apple)? Presumably not. But we do want to keep proper names like Bush distinct from the vocabulary word bush, and acronyms like CAD (computer-assisted design) separate from cad.

(ii) British and American spellings: colour and color. For some purposes (e.g. when analysing texts for meaning), we might want to consider these as the same type, but
on the other hand, it is often important to know that the text was written or published in the USA rather than in Britain, so we would want to keep them as different types.

b) this way of defining a ‘word’ only works for written language, not for spoken language

If you listen to spoken language carefully, we do not consistently speak in units of one word at a time, but in variable units, often in ‘tone units’ (also called intonation units, or breath groups), with pauses in between to breathe, to allow for a response, or for emphasis.

Also, Singleton points out (2000:7) that ‘both in the history of human language and in the development of the individual’, spoken language comes before written language.

Task 1.8
Using spelling to define a word is not so much of a problem, if we are only dealing with written English. But we are teaching spoken English as well.
What are the problems with transcribing spoken language as if it were written? Reflect on this for a few minutes before you read on.

In general, corpus creators do transcribe spoken language as written. But they face the problem of choosing between ‘accuracy’ and ‘normalization’. For example, are you going to transcribe as gonna or going to; if you choose the first, you won’t be able to compare written and spoken data easily, as gonna is much less frequent in written texts; if you choose going to, you are not representing the spoken text faithfully.

Spoken grammar can be reasonably described from corpus data, but a lot can be lost: for example, pronunciation, intonation, emphasis, pragmatics.

On the other hand, accurate encoding is very expensive (as it requires careful and laborious work by highly trained people), and makes the analysis (especially comparisons between written and spoken language) much more complicated.

1.4.2 Defining ‘word’ by Sound

Singleton distinguishes two approaches to defining ‘word’ by sound:

(a) the phonetic approach (which is language-specific, so we can restrict ourselves to English): the problem is that words are not separated by pauses in speech. Words are often run together (e.g. the written sequence ‘an apple’ becomes the spoken sequence /anapple/).

(b) the phonological approach (which is actually non-language specific; but for our purposes, we can look just at English): for English, we can use stress patterns, because each word tends to have only one (primary-) stressed syllable. But yet again, this is only approximate, a tendency, and is not consistent. For example, grammatical words don’t normally take stress, and fixed phrases/compounds have only one stress.

1.4.3 Defining ‘word’ by Meaning

(a) every ‘word’ is not always a unit of meaning. I know that I may have talked about the ‘meanings of words’ earlier in this Unit, but this was just for convenience. Indeed, lexicography depends on this convention. However, if I just say ‘bank’, you won’t know whether I mean ‘the side of a river’ or ‘a financial institution’ or ‘a building in which a financial institution is located’. Hence Hanks (2000) talks about the ‘meaning potential’ of words, and the ‘potential meanings’ of words.

(b) In general, meaning arises from context. I once attended a seminar given by John Sinclair entitled ‘Uniguity’. He argued that there is hardly any real occurrence of ambiguity, if one looks at an appropriate length of context.

(c) we recognize grammatical items such as the, and, of as words, but can we really say that they have a meaning?

(b) meanings are often associated with several words used together in specific sequences or combinations, rather than with individual words (e.g. fire engine, table tennis, civil service, take a look, have a bath, make up, at the end of the day, for the first time, spill the beans, etc).

(c) meaning can also be generated by units smaller than a word – morphemes: the -ly
at the end of angrily tells us that something was done or said in an angry manner, and the –ed at the end of walked tells us that we are talking about something that happened in the past. (I do not propose to discuss morphology – also called word-formation - in detail in this module. If you are particularly interested in morphology, you should read Singleton (2000) Chapter 3.)

We will look at ‘meaning’ again from various different perspectives throughout this module.

1.4.4 Defining ‘word’ by Grammar

The definition of a word in grammatical terms is that it should possess ‘internal stability’ and ‘positional mobility’. But ‘grammar words’ (such as the, and, of) are positionally restricted, so would be excluded by this definition.

Task 1.9
What are the problems with these definitions of word as applied to the mother-tongue of your students?

Students often use word-for-word translation equivalents in the early stages of learning English. To what extent do you think this helps, and to what extent might this actually hinder their progress?

Despite all the problems listed in sections 1.4.1 to 1.4.4, the ‘word’ is still a convenient and practical unit for many purposes, especially for corpus analysis. For any computer-based study of English, we have to use a definition of word that the computer will find easy to handle, so we opt for ‘any sequence of characters separated by spaces’. As we have seen, this ‘orthographic’ definition is highly problematical, but it can be gradually refined to suit our linguistic purposes, and allows us to use the power of the computer to look at more complex features of English using huge
amounts of data. We will look at this in more detail in Unit 2.

This may also be an appropriate moment to point out that ‘word’ is not the only problematic unit in language studies. Every other unit (morpheme, group, clause, sentence, paragraph, text, discourse) has its own definition problems, some of which will be discussed in later Units.

Now let us turn to an issue that is more directly related to the language classroom.

1.5 What does ‘knowing/learning a word’ entail?

Task 1.10: Read Singleton (2000) Chapters 9 & 10 for more on this topic.

This is obviously a crucial question for any language teacher, yet it is not at all easy to arrive at a sensible answer.

One of the problems, as Willis (2003:1-2) says, is:

‘What is “taught” is often not learnt, and learners often “learn” things which have not been taught at all. Learners often produce sentences such as: I am student or My father is engineer even though they have never been taught this, and even though their conscientious teacher is at pains to point out that the indefinite article is required here: You are a student; Your father is an engineer.’

There is also an important distinction between L1 and L2 learning. Singleton (2000:211) points out that:

‘a major focus of teaching as far as the first language lexicon is concerned is on enabling children to read and write words which they are already able to understand and use in speech…. With regard to classrooms where languages other than the first language are being taught, the typical scenario in this case is that when learners begin to be exposed to a second language in a formal educational setting, all aspects of the words they encounter are new to them.’

Willis (2003:17) issues a similar warning:
‘we should be careful not to overestimate the similarities between first and second language learning’.
He contrasts the L1 child who
‘puts together a string of lexical words supported by gesture and context, and depends on the adult’s willingness to work out meanings and act on them’
with the adult L2 learner, who
‘already speaks one language fluently and is able to use that language as a resource to help with the learning of a new language’.
Yet few adult L2 learners achieve native-speaker proficiency, whereas almost all L1 children do. Willis concludes that ‘there must be marked differences in the learning processes’.

Task 1.11: How do you assess when a learner ‘knows a word’ or ‘has learned a word’?

Roe (1998:Unit 3:24) reports that teachers say ‘a learner has learned a word when they can use it productively in normal communication’, and asks how teachers know when this has been achieved. Roe offers a ‘discrete points test’ and suggests that learners have learned a word when:

a) you recognize it when they say it in isolation

b) they spell it correctly when you say it

c) they replace it correctly in a gapped sentence

d) they can produce an appropriate dictionary definition

e) they can produce an appropriate translation

f) they can list its inflectional variants

These points can all be discretely tested and marked, but Roe recognizes that such tests are very intellectual and not psychological. In other words, they may still only reflect the learners’ knowledge about the language rather than their acquisition of the language as such.

Roe also reminds us of another distinction: between receptive and productive learning (op.cit:25, citing Haycraft 1978):
‘receptive vocabulary consists of words that the student recognizes and understands when met in context; productive vocabulary consists of words which student understands, can pronounce correctly, and use constructively in speech and writing.’

This topic will of course be taken up in much more detail in Unit 10.

Having looked at various aspects of lexis, let us now look at some aspects of grammar.

1.6 What does ‘grammar’ mean?

In Unit 7 of this module, Tom Bloor refers to the fact that ‘grammar’ can have many meanings, for example:

a) the properties of a language (e.g. its sounds, its writing system)

b) the way a language is organised (e.g. the allowed combinations of sounds, the structures that the words can occur in)

c) an account of a language’s properties/organization by a grammarian (e.g. Halliday)

d) the internal language knowledge of a speaker (e.g. you or me)

e) language-independent grammar (i.e. properties that all languages share)

f) a book on one of the above topics (e.g. Halliday’s Introduction to Functional Grammar)

Bloor suggests that we mainly use the first two meanings.

Willis (2003:28) says that he used to think that ‘grammar was about sentences and lexis was about words’.

Task 1.12
What did the term ‘grammar’ mean to you when you were at school?
What does it mean to you now as a language teacher?
What do you think it means to a researcher interested in SLA?

1.7 Different kinds of Grammar

Task 1.13
Re-read Chapters 1 & 2 of Willis (2003) carefully at this point, and do the tasks provided there, before moving on to the next section of the module. This will help you to explore the immediate and practical links between grammar and language teaching.

1.7.1 Singleton (2000:17-32) outlines various approaches to grammar that have some basis in lexis and the grammatical behaviour of individual lexical items:

a) Colligation – this is a term coined by Firth (see 1.1.3 above), which initially referred to ‘the co-occurrence of grammatical choices’ (Sinclair 1996, quoting Firth 1957b), that is sequences of word-classes or categories (as distinct from collocation, the co-occurrence of individual lexical items; see Unit 6, section 6.2), for example VERB + PREPOSITION (e.g. look at, refer to), or NOUN + TO-INFINITIVE (e.g. the right to do something, the decision to do something). However, even Singleton’s examples show that many people now use the term less rigorously and include combinations of lexical items with word-classes or categories (e.g. will + VERB, rather than ‘MODAL + VERB’; regret + VERBing, rather than ‘VERB + VERBing’).

b) Computational Linguistics – In general, computational linguistics is involved in trying to replicate and automate human language processes, in such fields as Information Retrieval, Automatic Summarizing, Speech and Text Generation, and Machine Translation.

Singleton, like many others, includes corpus linguistics as a branch of computational linguistics. But some computational linguists are much more ‘system-oriented’ and try to impose a ‘system’ on language texts, or on their procedures and applications. Whereas most corpus linguists are trying to glimpse and describe features of the ‘system’ through copious examples of ‘use’. Some people typify the latter approach as ‘bottom-up’, and the former approach as ‘top-down’.

For example, the Cobuild Grammar (1990) was created out of the ‘bottom-up’ lexicographic efforts that went into the Cobuild Dictionary, by looking at the grammatical behaviour of the thousands of individual words in the corpus and thence the dictionary, and grouping the words together in classes or categories according to the patterns of grammatical behaviour that were shared.

However, the insights gained by corpus linguistics are now being incorporated into many projects in computational linguistics.

c) Head-driven Phrase Structure Grammar – based on the notion that the grammatical preferences of the ‘head’ (or most important word) of a noun phrase or verb phrase helps to determine the grammar of the sentence in which the phrase occurs.

d) The ‘London School’ – the Firth, Halliday, Sinclair, and Cobuild tradition, already described in some detail in section 1.1.3, and mentioned at several other points in this Unit.

e) Valency Grammar – focuses on the grammatical preferences of individual verbs.

f) Dependency Grammar – extends the Valency Grammar approach to other word-classes.

g) Lexical-Functional grammar – an offshoot of Chomskyan grammar, which
acknowledges that lexical items have an impact on their syntactic environment

h) Chomskyan grammar – Government and Binding, the Projection Principle, Universal Grammar, principles and parameters, the Minimalist Programme.

I do not propose to deal with Chomskyan grammar in any detail here, as we will be focussing on Hallidayan grammar (Systemic Functional Grammar) in Units 7-9.

The most extreme version of Chomskyan grammar is similar to the historical linguistic tradition described in section 1.1.2, because in it ‘words were considered to be merely the observable elements through which syntax manifested itself’ (Singleton 2000:23-24). For more details, see Singleton (2000:23-28). The influence of Chomsky was extremely powerful in linguistics from the 1960s to the 1980s, and his current thinking is evident in a recent interview (Andor 2004) which is (at the time of writing) available on the Web.

1.7.2 Larsen-Freeman (2001:34) explores the multiple meanings of the term “grammar”: 'It is used to refer both to language users' subconscious internal system and to linguists' attempts to codify - or describe- that system' (my underlining).

The first meaning would include the L2 learner’s developing inter-language system. The second refers to what we will refer to as ‘descriptive’ grammars. Referring to the latter, she goes on to distinguish between Formal grammars (which take form or structure as their starting point) and Functional grammars (which start from language use and communicative purpose in social interaction – i.e. pragmatics).

She also looks at grammar in language education, where she touches on SLA research. It is certainly worth reading her short paper, which is (at the time of writing) available on the Web.

1.7.3 The Cobuild English Grammar (1990) is a descriptive grammar that was drawn up as the result of the analysis of a corpus of over 100 million words, but it also serves as a reference grammar for teachers and learners alike. The book contains a Chart (pp. xxiv-xxv) describing the way in which it is organised, which shows the following elements of grammar, each of which it deals with to some extent:

Task 1.14
Which model of grammar is used in the curriculum your students follow?
What elements in the list above does it cover, and to what extent?
What are the problems you have encountered in using that model?

With so many varying approaches to grammar, it is not surprising that we find it difficult to get an adequate overview. In Unit 2, we will look more closely at corpus analysis and the way in which that has affected traditional views of grammar.

We will return to a discussion of the relationship between lexis and grammar in Unit 6 when we look at multi-word units, before embarking on a more detailed investigation of one model of grammar (Systemic Functional Grammar or Hallidayan Grammar) in Units 7-9.


Andor J. 2004. The master and his performance: An Interview with Noam Chomsky. Intercultural Pragmatics 1-1:93-111
(also available at

Carter R & McCarthy M (eds). 1988. Vocabulary and Language Teaching. London: Longman

Cowie AP. 2000. The EFL Dictionary Pioneers and their Legacies. Kernerman Dictionary News, No 8, July
(also available at

Firth JR. 1957a. Modes of Meaning. In Papers in Linguistics 1934-51, London: Oxford University Press

Firth JR. 1957b. A synopsis of linguistic theory 1930-1955. In Studies in Linguistic Analysis, Special volume of the Philological Society, Oxford. 168-205

Halliday MAK. 1966. Lexis as a linguistic level. In Bazell CE, Catford JC, Halliday MAK, & Robins RH (eds). In Memory of J.R. Firth. London: Longman

Hanks P. 2000. Do Word Meanings Exist? Computers and the Humanities 34:205-15 (also available at

Haycraft J. 1978. An Introduction to English Language Teaching. London: Longman

Larsen-Freeman D. 2001. Grammar. In Carter R & Nunan D (eds). The Cambridge Guide to TESOL. Cambridge: CUP
(also available at

Lewis M. 1993. The Lexical Approach. Hove: LTP

Owen C. 1993. Corpus-Based Grammar and the Heineken Effect: Lexico-grammatical Description for Language Learners. Applied Linguistics Vol. 14, No. 2

Robins RH. 1964. General Linguistics. An Introductory Survey. Harlow: Longman, Green

Roe P. 1998. Lexical Studies. Aston Msc Module

Sinclair JM. 1966. Beginning the Study of Lexis. In Bazell CE, Catford JC,
Halliday MAK & Robins RH (eds). In Memory of J.R. Firth. London: Longman

Sinclair JM, Jones S & Daley R. 1970. English Lexical Studies. Published (2004) as
Krishnamurthy R (ed). English Collocation Studies. London and New York: Continuum

Sinclair JM (ed). 1987. Looking Up - An account of the COBUILD Project in
lexical computing. London: HarperCollins

Sinclair JM. 1991. Corpus, Concordance, Collocation. Oxford: Oxford University Press

Sinclair JM. 1996. The Search for Units of Meaning. Textus IX:75-106

Singleton D. 2000. Language and the Lexicon. London: Arnold, and New York: OUP

Wilkins D. 1972. Linguistics and Language Teaching. London: Edward Arnold

Willis D. 1990. The Lexical Syllabus: A New Approach to Language Teaching. London: HarperCollins

Willis D. 2003. Rules, Patterns and Words. Grammar and Lexis in English Language Teaching. Cambridge: CUP


Willis D. 2003. Rules, Patterns and Words. Grammar and Lexis in English Language Teaching. Cambridge: CUP [Chapters 1 & 2, plus tasks]

Singleton D. 2000. Language and the Lexicon. London: Arnold and New York: OUP [pp 1-10]

Waring R. 2002. Basic Principles and Practice in Vocabulary Instruction.
(available at

Waring R & Nation P. 2004. Second Language Reading and Incidental Vocabulary Learning. Angles on The English-Speaking World Vol 4
(also available at

Larsen-Freeman D. 2001. Grammar. In Carter R & Nunan D (eds). The Cambridge Guide to TESOL. Cambridge: CUP
(also available at


Andor J. 2004. The master and his performance: An Interview with Noam Chomsky. Intercultural Pragmatics 1-1:93-111
(also available at

Cowie AP. 2000. The EFL Dictionary Pioneers and their Legacies. Kernerman Dictionary News, No 8, July
(also available at

Hanks P. 2000. Do Word Meanings Exist? Computers and the Humanities 34:205-15 (also available at

Meehan, P. (2003). Lexis – the new grammar?
(available at