SLA Topics     SLA Bibliography    Vivian Cook   

Vivian Cook

(the basis for a chapter in L2 Writing Systems, 2005)

Since the decline of audiolingualism, there has been little public debate about the respective roles of spoken language and written language in language teaching. This paper argues that it is time to start thinking about this relationship again. The basic question it asks is how teaching should reflect speech and writing in the early stages of language learning, concentrating on the beginners level rather than more advanced levels of writing, which have received more attention.

The priority of speech has been a constant theme in teaching methodology (Banathy & Sawyer, 1969), forming article 1 of the International Phonetics Association in the 1880s 'Foreign language study should begin with the spoken language of everyday life' (cited in Stern, 1983). The audiolingualism of the 1960s took the first principle of ‘scientific language teaching’ as ‘Speech before writing’ (Lado, 1964). The only twentieth century exception was perhaps the short-lived Direct Reading Method (Coleman, 1929).

The bias towards speech has not altered with the decline of audiolingualism. The Natural Approach to language teaching emphasises ‘the ability to understand and speak’ (Terrell, Egasse & Andrade, 1990, p.xix). The English curriculum in Cuba insists on 'The principle of the primacy of spoken language' (Cuban Ministry of Education, 1999). Communicative language teaching uses ‘class-room activities designed to get learners to speak and listen to each other’ (Scrivenor, 1994, p.62). Alternative methods such as Total Physical Response (TPR) use story-telling, not story-reading (Seely & Romijn, 1995). Few people have taken the opposing view that written language should be the basis of second language (L2) teaching, except, perhaps, for Krashen's advocacy of Free Voluntary Reading as a form of comprehensible input (Krashen, 1993). In 1984 Howatt claimed ‘The spoken language for example is promoted with more determination now than at any time since the Reform Movement’ (Howatt, 1984, p.289), which seems as true two decades later.

As always, the discussion is clouded by the gap between 'official' language methodology and the reality in the classroom; what relationship do the recommendations by methodologists, syllabuses etc have to what actually goes on in the classroom? That is to say, even if teachers are exhorted to use the spoken language, do they really do so? The bridge that is used here is course-books; if teachers' practice departed too much from the coursebook, then presumably it would not be on sale for very long.

To show how written language is handled in teaching, let us look at some beginners' course-books available for international markets, not just for English but across other languages, namely seven adult low-level courses: five for English – Atlas (Nunan, 1995), Reward (Greenall, 1994), Changes (Richards 1998), Tapestry 1 Listening and Speaking (Benz & Dworak, 2000), Headway Elementary (Soars & Soars, 1993) – and two for other languages – Ci Siamo (Guarnaccio & Guarnaccio, 1997) for Italian and Libre Echange (Courtillon & de Salins, 1995) for French. Changes, Libre Echange and Atlas claim to be suitable for beginners, Reward for 'intermediate', Headway for 'elementary' and Tapestry for 'high beginning'; Ci Siamo does not specify level. All the courses apart from Ci Siamo are monolingual and all except for Tapestry are stand-alone volumes. These are believed to be representative of 1990s language teaching materials. The usual convention will be adopted here of angle brackets for written forms such as <bl>, parallel to the use of slants for phonetic script /bl/, and italics for examples such as table.

Written transcripts of spoken language form an essential component in most courses, usually highly scripted: Cameriere: Buonasera. S'accomodi. Ecco il menù. …(Ci Siamo). Tapestry 1, Reward and Headway Elementary provide scripts of the listening materials at the back of the book. Changes has two short scripted conversations in every unit. Ci Siamo uses techniques from the teen magazines of photo-stories with captions and speech-balloons. Most scripts in course-books resemble plays. Indeed Libre Echange uses authentic film scripts from inter alia La Dentellière and Pauline à la plage. Though the early Reform Movement valued phonetic script as a way of showing spoken language on the page (Sweet, 1899), this practice survives only in the phonetic symbols used in some vocabulary lists (Headway) and sparingly in pronunciation practice in Libre Echange and Reward.

The use of scripts shows written language in the service of speech. The written text provides a permanent form of the speech the students encounter that can be referred to, analysed and revised. Scripts are not teaching writing since the characteristics are all supposedly those of spoken language. They provide incidental exposure to the characteristics of written language but in an arbitrary and unplanned way; for instance leçon <<zéro>> of Libre Echange introduces three conventions of written French in its title (<ç>, <é> and 'goosefeet' quotation marks << >>). The reverse concept of written language to be read aloud is never employed in these coursebooks, even if familiar from news-readers or priests reading texts aloud.

Written descriptions of pronunciation or grammar are given in the coursebooks, sometimes in the L1: Regular verbs ending in -ire are divided into two groups … (Ci Siamo), though mostly in the L2: Le «u» de «tu» tombe devant une voyelle (Libre Echange), extended to learning strategies in Tapestry 1: Use music to help you learn English and to improve your pronunciation. Related uses are displays of grammar in substitution tables and word-sorting: Write the words in the correct columns. Wednesday night Friday singer … (Atlas). Many courses display structures in columns as the main means of 'explaining' grammar, whether 'going to' future or 'have got' (Changes), as described in Cook, (1989).

Written explanations again exploit the permanency of written language, columns its visual display properties. Perhaps only in a coursebook is grammar explained in this way. (And furthermore they embody a pre-1957 view of phrase structure (Cook,, 1989)).

• conveying meaning

Written language conveys the meaning of unknown words, sometimes in the L2, sometimes in the L1. One possibility is to label pictures – parts of the anatomy of Michelangelo's David: Il corpo 1) i capelli 2) la fronte…: (Ci Siamo) or: Write the name of the correct symptom under each picture…: a fever, a stuffy nose, a headache … (Tapestry). Another is lists of translation equivalents; Al Bar. il caffè coffee il cameriere waiter … . (Ci Siamo). This overlaps with the practice of matching pictures with descriptions: Match a picture with an activity. Dancing, visiting museums, doing crosswords, …: (Headway), or of identifying elements in a photo: Which of the things can you see in the photo? Apples, bananas, beef, beer … (Reward).

Outside coursebooks, diagrams are used in manuals for car maintenance or drawings of cooking utensils, in which the point of the labelling is to aid the activity; it doesn't matter whether you know what a gudgeon pin is the day after servicing your car provided you have oiled it correctly. In course-books, however, the point of the diagram is the link between the image and the new word.

• giving instructions about exercises

Instructions for exercises are given in writing, sometimes in the L2: Comparez les differents formulations (Libre Echange), sometimes in the L1: Interview a classmate and find out how they spent last weekend (Ci Siamo); Changes gives the script for a Total Physical Response sequence (a) Close your book, please. (b) Open your exercise book, please. (c) …. , illustrated by drawings of the actions. Possibly such instructions are more for the teacher than for the students as their language is often way above the students' level: Lesson 1 of Atlas for example has: Match the countries and nationalities and then say the words, involving not only vocabulary which has not been taught, for example words, but also imperatives match/say and structures co-ordinated by and – Match … and … say.

Outside the classroom instructions are a restricted genre of written language. Written instructions mostly tell individuals how to do a tasks such as filling in their income tax return or baking a cake, not how to interact with a partner or a group of people in the manner of most classroom instructions.

• asking comprehension questions

The text of the spoken conversations is often followed by written questions checking comprehension: Fabienne, Carmen et Eric sortent: ou vont-ils? (Libre Echange). The answers may be ticks: Who owns these things? Listen and tick the right name (Changes), multiple choice questions(Tapestry), matching complaints with the advice given by a doctor (Reward), or filling in a chart such as: Listen to four people talking about the food they eat on different days of the week. Complete the chart. (Changes).

Once again written questions about what you have heard orally would rarely occur outside a classroom. However realistic the spoken language that students hear, subsequent comprehension questions are not a typical use of written language.

Many spoken exercises rely on fragments of written language as props: could the communicative method indeed exist without written language prompts?

• lists.

Often these props are lists of words, for instance: Match the foods and drinks to the words in the chart (Changes). Sometimes they are prompts for ideas: Computers. Can they ….? Count smell translate … (Headway), or for sentences: Make a sentence with each of these weather words … rain dry sunny … (Tapestry). A variant is sorting word lists into columns or charts. Reward makes students sort ballet cinema film interval ticket into What's on?, Where and Related words.

Little outside the educational context provides a model or a purpose for these isolated pieces of written language. The only parallel might be making a speech from notes, which is self-prepared and far more complex.

• realia

Information necessary for the exercise can be conveyed through realia and graphics, such as the labelling on charts, horoscopes, calendars – both Ci Siamo and Libre Echange have saints' calendars – shopping catalogues: i jeans 1 avoria L.50.00, 2 nero L.79.00 … (Ci Siamo), and ephemera such as supermarket fliers (Atlas, Ci Siamo). Maps are particularly popular whether San Francisco (Atlas), West Valley College Saratoga (Tapestry), London (Changes), Dublin (Reward) or Urbania (Ci Siamo). Occasionally the coursebooks provide actual statistical charts, for example, regions of Italy popular with tourists (Ci Siamo), percentage of U.S. population that was foreign-born (Tapestry) and les professions des Français (Libre Echange).

While these represent a normal use of written language for display and information, the purposes for which the students use them may be less authentic, partly because they are learning, say, the names for the clothes rather than deciding what to buy, i.e. a codebreaking rather than a decoding activity (Cook,, 2001b). The written language consists of noun phrases, proper names, prices etc rather than being 'full' sentences or having textual coherence – block language.

• fill-in forms and charts

Students fill in copious amounts of information in invitations, diary planners: Daily Planner Monday Morning …… Afternoon …. Evening … … (Atlas) and forms: Carte Internationale d'Embarquement 1. M/Mme/Mlle ___________ Nom ______ … (Libre Echange). Mostly their response is a single word or phrase rather than a complete sentence or paragraph. This shades over into the innumerable charts for students to fill in, already mentioned. In Changes for instance the listening activities often rely on a chart such as: Listen to four people talking about the food they eat on different days of the week. Complete the chart.

In everyday life we do of course have to fill in such forms from time to time but mostly this is an unavoidable chore. Sorting information into columns manually is rare – would anyone not use their computer for this now? The language is fragments and isolated words. Most of the charts and forms have no outcome other than providing material for a teaching exercise.

• sentence completion

Students fill in blanks in sentences before saying them aloud: Et …… laitues, il y a … … laitues (Libre Echange), or they construct sentences from jumbled words: Rearrange the words to make questions and answers and then practice them: you/where/live do? ... (Atlas). Sometimes the missing item is a word: Ana is Jonathan's ___ (Changes), sometimes a longer phrase Sure. ___ Campus Center Walk ___ the Campus Center (Tapestry).

In a sense this exploits the permanency of written language so that the whole sentence can be present simultaneously, again using written language in a way seldom encountered outside a classroom.

• making up sentences

Students are given questions and answers in a jumbled order and have to pair them appropriately, for example My perfect weekend as seen by Stephen from Leeds and Paula from Nottingham (Reward). Sentences are also constructed from jumbled words: I dessert she want any steak don't fries they coffee some he wants (Atlas). A list of words provides a skeleton for retelling a story the students have heard: accident street break down capital cost … (Reward). Or indeed a sentence can be directly called for: Write an opening sentence about what happened to Bob (Reward).

This too is a purely pedagogical use of written language, inconceivable outside a classroom, relying similarly on reading aloud.

Most course-books also present written language through continuous texts longer than a single sentence.

• short information texts

Cultural information is often presented through short texts: La France au Quotidien: les palmarès des villes françaises … (Libre Echange) or magazine-style written interviews: Intervistiamo una nota stilista italiana … (Ci Siamo). Short texts are often used as a basis for exercises, typically about 50 words long: biographies – Betty aged 95 from Nottingham (Changes), – first person accounts – Sun Hee Shi talks about her birthday (Changes) – or factual accounts – Amazing Animals (Did you know the kangaroo can't walk at all?) (Changes). Longer texts around 400 words are found on Superman (Tapestry 1), and The Day of the Dead (a Mexican festival rather than a 50s horror film) (Reward). Only Reward features a real short story, by Roald Dahl, and extracts from books such as Paul Theroux's The Kingdom by the Sea.

These short texts resemble pieces of travel guides or children's textbooks. Little of our everyday reading consists of passages of this type (particularly when they are pretexts for arranging information in columns or completing sentences).

• letters

Specimens of letters occasionally form a basis for the students' own writing, for instance refusing an invitation (Atlas) or telling a friend about Italian pastimes (Ci Siamo).

Continuous written texts other than scripts of speech or instructions for exercises are in fact sparse in some course-books, though only Atlas appears to shun them apart from a few notices and scraps of newspapers. Virtually all the courses use short texts whose similarity to anything outside the coursebook is minimal; they do not have the display or structure of newspapers, say, but are straight pieces of information to use for teaching exercises.

Let us then sum up the conclusions about course-books reached in this section:

• written scripts are primarily a way of teaching spoken language by providing a permanent record, not of teaching written language.

• written language is often a device for explaining, giving instructions etc, i.e. a kind of meta-language of teaching, rather than a way-in to writing itself.

• written language within teaching activities is mostly a pretext for spoken exercises, involving uses of language not encountered outside textbooks.

• texts are mostly restricted to short quasi-factual biographies etc with some longer texts about 'interesting facts', again seldom recognisable as text types that would occur outside a teaching context.

These features seem typical of the course-books. Though they may be interpreted differently by different teachers, there is no reason to think that the average beginning student will encounter a totally different range of written language than that represented by these coursebooks.

The next step is to review some of the fundamentals of written language, dealt with at greater length in Cook, (2001a, to appear); section 4 will develop the teaching implications.

The 'standard' model of reading aloud relates letters to sounds via two mental routes (e.g. Paap et al, 1992). One 'sound-based' route uses correspondence rules between speech sounds and written forms: an English reader sees <tack> and consults mental rules that <t> corresponds to /t/, <a> to /¾ /, and <ck> to /k/. This route relies on rules for converting letters to sounds and vice versa – ‘assembling phonology from a word’s component letters’ (Katz & Frost, 1992, p.71).

The other 'visual' route utilises direct links between written forms and meanings. corresponds in Chinese to the meaning 'person' regardless of its different spoken forms in the various Chinese dialects. The English reader sees <colonel> and consults a mental list of words to establish that it corresponds to /k‘ nlÛ /. This route enables reading to be silent so that readers can process written language at speeds much faster than they can read aloud.

These two overall routes not only distinguish the writing systems used in different languages, but also operate within an individual user of a single language. Frequent English words are accessed as visual instances (Seidenberg, 1992); <the> for instance is perceived as a whole word rather than converted into sounds letter by letter; if you ask native speakers to cross out <e>s in a passage of English, the <e> in <the> is invisible (Cook,, to appear).

So written language involves not just the conversion of sounds into writing and vice versa but also direct connections to meaning. The assumption that sounds connect only to letters is a product of our own cultural bias. An L2 student who is moving from Chinese to English has different problems with written language from a student who is going from English to Spanish. L2 learners transfer their preferred route to the new language; thus Chinese learners of Japanese differ from English learners in being more visually oriented (Chikamatsu, 1996; Holm & Dodd, 1996); advanced Chinese learners on the other hand remain slower at reading English than Spanish learners (Haynes & Carr, 1990). Literate students bring to L2 learning the ideas about language created by the writing system of their first language, for good or for ill.

Course-books assume that written language simply provides a transcription of speech. This neglects the differences between languages, in particular the difficulty in acquiring a script of a different type, say English speakers learning Japanese or Chinese speakers learning Spanish. A single language of one overall type also contains elements of both systems: English is not purely sound-based, Chinese not purely meaning-based. A language user relies on both the sounds route and the visual route, to a greater or lesser degree depending on the language concerned and on word frequency.

The implications of the two routes are that teaching needs to acknowledge them rather than ignore them. The coursebooks do not deal overtly with the two routes. Students from different overall L1 writing systems are unprepared for the new system they are encountering, leading to persistent problems for Chinese students. The students, and probably the teachers, have no awareness that the two routes actually exist. Implicitly this reduces the teaching of English writing to apologies for its lack of system. Advanced L2 students still write phisical, neccessary and new (knew) because they have not been taught to treat some words as individual visual items (mistakes come from The 1400 corpus (Cook,, on-line)), and feever, persent and colledge because they believe writing is the reflection of speech. Some attention to both routes can do nothing but help.

b) the characteristics of written language

Here we will list some of the properties of written language briefly; other lists can be found in Brookes & Grundy (1998), Bygate (1998) and Ur (1996).

• vocabulary. Some words exist primarily in the written form, say notices on road verges No parking on the greensward or on train doors Alight other side.

• grammar. Written language has a proportion of about 1 grammatical word to 1.2 content words compared to the 1 to 0.6 of speech, leading to the so-called 'lexical density' of written language (Halliday, 1985). Written language is more contextualised, making more explicit deictic connections to the situation than speech by naming people, times and places: Miss Scarlett killed Professor Plum with the dagger in the library last Thursday, not She killed him with it over there yesterday.

• punctuation. The contrasts conveyed through punctuation in written language, <Joan!> versus <Joan?>, are similar but not identical to those conveyed with intonational pitch changes. Punctuation has own unique features, for instance the use of the semi-colon as a grammatical item in sentences such as She told him the truth; the others lied (Nunberg, 1990).

• distinctive elements of writing. Written language uses visual distinctions not available in speech, for example capital letters <NO ENTRY> and word spaces <Howdoyoudo?>. The conventions of written English differ from those in other languages, and there are differences between dialects, such as capitals after colons in American English but not British English.

• spelling rules. While newspaper pundits and spelling reformers bemoan the inadequacies of English spelling, others point to the subtleties of its organisation. Carol Chomsky (1970) sees 'silent letters' in words such as <sign> as reflecting the underlying lexical form seen in <signature, assignation> etc. There are also rules and patterns to English spelling, such as the rule that only function words can have less than three letters (Carney, 1994), leading to pairs such as <I/eye>, <be/bee>, <in/inn> and many others. Far from being unsystematic, English spelling has a complex set of systems, Albrow (1970) for example distinguishing basic, romance and exotic systems.

These linguistic characteristics of writing have been ignored in teaching. The written language of course-books does not reflect the standard properties of lexical density nor a proper range of punctuation: how could it when so much is sentence fragments? Far from being more contextualised, the written language of teaching is a limbo world of charts, lists, short texts, explanations and instructions.

This bias might be excusable if specific attention were paid to the actual rules of writing. Yet nowhere in the coursebooks can one find discussion or presentation, say, of spelling issues. If students were taught basic rules of English, they would not make mistakes such as usefull, ocured and dissappointment (rules of consonant doubling – 18% of the corpus) or studys, alwais or joyning (rules of <i/y> alternation); if they had been warned about transferring elements of their own L1 pronunciation through sound-letter rules, Japanese might not write grobal and literery (literally), Greeks Grade Britain and Gambridge, Arabic speakers punishement and subejects.

• permanency of writing. Speech is fleeting: writing permanent. A written letter is present as long as, or as often as, we want to look at it; a spoken phoneme is gone as soon as it has been uttered and can only be checked against our internal memories or via an electronic recording. Speech is used for instant warnings Fire!, writing for permanent notices <KEEP OUT>. 'To write is indeed the only way of keeping or recapturing speech since speech denies itself as it gives itself' (Derrida, 1992). Writing is a memory substitute both for the society through written constitutions, history books, contracts, etc and for the individual through diaries, shopping lists, lecture notes and so on.

• first and final drafts. Normally readers see the final written draft, which may have been through many revisions and spelling checks; writers are able to go over and over their drafts. Speech is composed on the wing; there is no second chance to get it right. This division is not, however, integral to the speech versus writing dichotomy; some written media such as e-mails and text-messaging now use first drafts; some spoken media use final drafts, witnessed by the decline in truly 'live' television programmes other than sport.

These characteristics are seldom seen in the course-books. Students may employ learning strategies based on the written language, whether making notes for future reference, keeping lists of vocabulary, or consulting dictionaries. Students, like L1 children, may need practice in developing a critical approach to successive drafts of their essays, letters, or indeed theses. A written sentence can always be improved; a spoken one cannot. All of these essential attributes of writing are relatively unexploited in the course-books.

Inevitably the differences between the two media have led to speech and writing being used for different purposes.

• status. Writing often has higher status than speech because its permanent form allows it to record laws, contracts etc. In modern societies 'serious' literature is written down as highly-wrought novels, not improvised as oral epics; poetry is found in books, not folksongs. In many societies, holy writings capture the very words of God. In most, perhaps all, literate societies, written language exerts influence over speech, for example the historical pronunciation changes from /f‰ rI d/ to /f˜ :hed/ for <forehead> in English and the preservation of the /a/ /a:/ contrast in French by the circumflex accent <tache>/<tâche> (Valdman, 1966). With the exception of linguists, most people regard writing as more important than speech; 'One could nearly say that in a "literate culture" speech is the spelling of writing' (Kress, 2000, p.18); a main component in Noam Chomsky's Minimalist Program is revealingly called 'Spell Out', not 'Speak Out' (Chomsky, 1995).

• social roles. Listeners and speakers are typically present in the same situation at the same time, know each other's social roles, interact with each other and exchange listener/speaker roles. Readers and writers are typically separated by time and place; indeed the writer may be long dead. A speaker is a visible, clearly identified, individual, a writer an invisible, even anonymous, person. The social relationship is indefinite; what is our social role to T.S. Eliot when we read Burnt Norton? In writing the speaker is not usually identified and the reader is not addressed directly.

Students should be exposed to the appropriate functions for both types of language, spoken language for the less formal, written language for the more formal, according to their particular needs. Some students may need informal roles – 'Hi, I'm Jim' – some more formal – 'Dr Livingstone, I presume'. The emphasis on informal or tourist conversation in beginners teaching has often not fitted the formal nature of the students' real-life encounters, let alone the written language they may need. They may require informal written language to use chat-lines and e-pal systems as well as the extra formality of many types of writing compared to speaking.

Becoming literate in itself has effects on the individual:

• literacy and cognition. Luria’s work showed that non-literate people reason in a less abstract way (Luria, 1976). People who have learnt to read are never the same again: they have different perceptions of the world. People who can read store information differently from those do not (Goody, 2000). The actual brains of literate people differ from those of non-literates (Petersen, Reis, Askelöf, Castro-Caldas & Ingvar, 2000).

• literacy and the spoken language. Adults’ perceptions of spoken language are influenced by their knowledge of writing. English people believe there are more sounds in ridge /r I d Z / than there are in rage /r eid Z / (Derwing, 1992) because of the extra letter <d> in the written form. One debate is whether the initial acquisition of L1 writing depends on prior knowledge of phonemes or, in reverse, knowledge of writing shapes the knowledge of speech into phonemes; <n> is frequently omitted by English children because they hear it as a nasalised quality of the preceding vowel rather than as a phoneme (Treiman, 1993). As Olson (1996, p.100) puts it, 'Writing systems create the categories in terms of which we become conscious of speech'. Written English is pre-analysed into words by spaces, types of nouns by capital letters and grammatical constructions by punctuation.

An L2 learner who already knows how to read thinks in different ways from a pre-literate child. Literate L2 learners have already established the two routes in their minds; they inevitably relate the new language to writing in a way that the non-literate do not; they are largely dependent on the type of writing system they have learnt in their first language. The changed mind of the L2 learner needs to be accommodated in teaching rather than ignored. The literate L2 learner cannot enter into language acquisition in the same way as the pre-literate native child. I can remember being surprised by the students in a strictly audiolingual class who were surreptitiously writing down the dialogues they heard under their desks. Pretending that students are not fully equipped with a knowledge of writing ignores the changes in the minds of literate L2 users and the powerful tool for learning provided by writing.

Perhaps, however teachers had good reason for assuming that the written language was just an adjunct to the spoken language. Let us then try to examine some of the reasons put forward for the priority of speech.

One overall justification is how children acquire their first language, just as it is for many nineteenth century teaching principles (Howatt, 1984; Banathy & Sawyer, 1969). As Harmer (1998, p.53) puts it, ‘Because many people acquire languages by hearing them first, many teachers prefer to expose students to the spoken form first’. If the only successful method of acquiring a language is that used by L1 children, the classroom should recreate the characteristics of the first language as closely as possible. Teaching cannot depend on things which are unavailable to L1 children, whether the ability to write, to understand grammar or to use another language.

This reflects more common-sense than theory or research. If the argument were true, no research would be needed into L2 acquisition or indeed into language teaching, since teaching needs simply to replicate the conditions of L1 acquisition. But L2 learning is intrinsically different from L1 learning because the L2 learner already knows how language works (Halliday, 1975) and how to impute mental intentions to other speakers (Tomasello, 1999). At best it might be said that in some respects, L2 learning is like L1 acquisition, in others it is not. The multi-competence viewpoint, however, denies that the comparison is proper since the L2 user has a complete system of two languages, not two separate systems (Cook,, 2002). There is no compelling reason why the accidental sequence of L1 acquisition should be relevant to literate L2 learners.

Most linguists and philosophers from Aristotle to Rousseau, from de Saussure to Bloomfield, have indeed believed in the primacy of speech, typified by Lyons (1968, p.38) 'the spoken language is primary and … writing is essentially a means of representing speech in another medium'. The arguments they present come down to the following:

• speech comes before writing historically. Writing came thousands of years after spoken language in the history of the human race. Indeed in one sense 'history' is distinguished from 'pre-history' by the existence of written records. However, a diachronic argument does not necessarily provide a synchronic explanation (Householder, 1971) and ignores the changes that written language makes to a society.

Even if the historical argument were true, its use would involve ontogeny repeating phylogeny, on a par with a modern diet consisting of wild grain and meat we have slaughtered ourselves. There is no intrinsic reason why the classroom should parallel earlier societies, despite Taeschner's (1991) amazing language teaching method in which children recreate a primitive tribe. The invention of writing in itself changed the society that children and L2 users fit into. Human beings inherit the cultural achievements of their ancestors (Tomasello, 1999), partly through written language. Silent reading. for example, once a difficult and unusual skill, is now taken for granted in the primary school, thanks to the cultural innovation of word spaces in the 8th century AD (Saenger, 1997). L2 teaching should no more take pre-literate societies into account than physics teaching should cover medieval alchemy.

• many languages lack a written form. Only within the last century did many languages acquire a written script, which some still lack. Conversely it is hard to imagine a language that exists in writing without having a spoken form. The lack of a written form for some languages is an interesting observation about the world and the spur for the work of the Wycliffe Bible Translators. It does not force teachers to stress speech, unless they are concerned with students from a non-literate society.

• many individuals cannot use written language. While it is hard to think of a non-disabled person who does not speak, many people do not learn to read and write, for whatever reason. The world-wide illiteracy rate for the year 2000 given in the UNESCO Statistic Yearbook (2000) is 20.6%, that is to say 876 million people cannot read and write. Mass literacy is a recent development in the world’s history. The fact that some individuals cannot write is crucial for teaching students who are not literate in their first language; it is, however, irrelevant for the rest. It is more to the point to consider which writing system they possess in their first language.

• children automatically learn to speak but have to be taught how to read and write. Talking is as normal for human beings as walking; writing is an acquired skill, like cycling or ice dancing. No normal child fails to learn to speak by interacting with its parents or caretaker; only a few children learn to read without being taught. The fact that L1 children need to be taught writing is not relevant to the teaching of literate L2 students.

There are perhaps three supplementary reasons that can be mentioned.

• students' needs. Students in some parts of the world may intend only to use the second language orally and never to open a magazine or browse a web-site. The obvious solution for them is no written language at all, as in a strict audiovisual method. Using written language as a prop for spoken is not justified.

• students' demands. Some students and teachers everywhere in the world may see the spoken language as their goal. They are part of the same culture that produced this attitude in methodology. Like other aspects of teaching such as the concept of the native speaker, it is hard for teachers to argue against popular beliefs ingrained into society and arguably, as servants rather than masters, it is up to teachers to fall in with their students' demands. Yet accepting this argument would on the one hand prevent any educational change, on the other fail to support the many students who want written language.

• interference. Traditionally teachers have felt that early writing forces the students into spelling pronunciations; since the English writing system is only partly sound-based, this creates inaccuracies and mislead students. Partly this reflects the assumption that English only has the sounds route available: if teachers and students were aware of the visual route, they would not be so concerned. Partly it ignores the interference that goes in the other direction: students may misread things because of the spoken form. Without the L2 written language, they may well be pushed back on their L1 written forms.

The position of writing in teaching has then to be related to how writing functions in the mind of an L2 user, to the variations in writing systems and to the purposes for which the L2 user may put the L2. While some writers disparage writing as being of negligible importance (‘Nowadays most people actually do very little writing in day-to-day life’ (Scrivenor, 1994, p.156)), many people now spend much of their working day at a computer keyboard and, since the decline of the secretary, spend a high proportion of their time receiving and composing letters, e-mails and text messages. English language education for non-English-speaking students is a continual series of hurdles of written essays, projects and examinations.

We can now draw the argument together to make some conclusions about the current and potential roles of written language in language teaching.

As we have seen, the major use of written language is to support the activities of the classroom. Written language has been seen as a teaching aid, as neutral as a blackboard or an overhead projector, not as having value for the students in their using and learning of language nor as itself presenting them with problems – say, recognition of words by students from L1s without word divisions, uses of capital letters, untaught in beginners' course-books, or indeed direction of writing whether left-to-right or right-to-left. The written language is the servant of teaching rather than having a role of its own; it is used as a tool, not as a type of language the students actually need.

Obviously teaching may subordinate written language to spoken language if it thinks this is in the students' interests. But this decision should be made after considering the distortions it imposes on the written language: is the price worth paying? It is also a matter of proportion; how much of the written language the students encounter is a prop for writing, how much writing for its own sake? Above all are the students' goals best served by sacrificing writing to speaking? While most students may indeed need spoken language, this hardly prepares them for the reading and writing tasks of everyday life, whether looking at the newspaper, reading a novel or writing an e-mail.

Time and again the previous discussion questioned whether typical teaching uses of written language existed outside the classroom. Opening one of these coursebooks at random, you see a page consisting of layouts of sentence fragments, lists of words, fill-in charts, etc. A typical page of Atlas, say page 57, has four exercises: exercise 1 is drawing lines between questions and answers; exercise 2 making up sentences from groups of words; exercise 3 filling in the blanks; exercise 4 writing 'a lot of' or 'enough' in the blanks in sentences. This represents an almost unique form of print, different from most books, magazines, newspapers or other everyday writing. The only similarities are with commercial publications for selling goods, say catalogues or travel agent's brochures, and with official forms for Income Tax returns etc. In our daily lives we seldom encounter lists of words that have to be sorted into groups, sentences with missing bits, written questions on our daily conversation, and so on.

Nor is the language that the students have to write any less fragmentary. Students fill in bits of noun phrases on forms; they complete sentences with words or phrases; they make graphic marks on paper. They seldom write a complete sentence, let alone a complete text; they rarely make up lists for themselves relevant to their learning or interests. The emphasis on speech has led to a misuse of writing. The teaching of the spoken language rarely distorts its nature to the same extent; at best the students use full sentences of the language in conversational turns, at worst appropriate sentence fragment answers to questions.

The solution for language teaching is then to teach the properties of the written language. One step in this direction is to use authentic written language, that is to say examples of writing produced for purposes other than language teaching, whether by native speakers or non-native speakers. At least then the teacher has the protection of knowing that the written language has its natural qualities. Another step is to systematically teach the relevant properties of language whether writing system, spelling or other conventions. As well as the use of written language for teaching, the students should at least encounter writing used as a part of human experience. None of this would be extraordinary when applied to speech; it is taken for granted that spoken language is presented as speech functions, conversations, interactions, sometimes authentic, sometimes simulated at a level the student can cope with. Only the bias towards speech has stopped the same criteria being applied to writing.

A vogue word in 1990s language teaching was task-based learning; indeed Atlas claims specifically to use 'task-based methodology'. Jane Willis (1996) lists six main type of task: listing, ordering and sorting, comparing, problem solving, sharing personal experience and creative. Atlas 1 (Nunan, 1995, teacher's book) uses ten types of task including predicting (for instance 'predicting what is to come on the learning process'), conversational patterns ('using expressions to start conversations and keep them going') and co-operating ('sharing ideas and learning with other students').

As we have seen, the tasks employ written language to foster speech. Take the sentence-invention exercise type we have often come across, typified by Tapestry: Make a sentence with each of these weather words. Practice saying the sentences to yourself. Then say the sentences with a partner. The sheer arbitrariness of sentence completion was something that the Reform Movement of the late nineteenth century attacked in previous methods: 'connected texts on worthwhile topics were clearly preferable to the pointless sentences of traditional grammar' (Howatt, 1984, 173). In its day the communicative teaching movement similarly reacted against the meaningless uncontextualised sentences of drills. Nevertheless sentence invention is a staple in nearly all the coursebooks. The lesson of the Reform Movements has not been applied to the teaching of written language. Spoken language in the classroom tries to be connected, relevant discourse; written language is by and large unconnected fragments acting as the impetus to exercises or as the source of explanations and instructions, largely true even of course-books specifically aimed at writing, such as Tapestry 1 Writing (Pike-Baky, 2000) which has a similar mixture of fill-in sentences, vocabulary lists and sentence fragments.

One unacknowledged element in many tasks is reading aloud. In some cases this is integral to the exercise, say fill-in sentences or substitution tables where the product is the student saying aloud what is written down. In other cases the exercise could not take place without some reading aloud, words in lists for instance. In so far as methodology books mention reading aloud, they warn against it: 'Do not ask students to read texts aloud' (Cross, 1992, p.80). Yet these coursebooks require students to read something aloud on nearly every page. While this is not quite the same as reading whole texts aloud, it is nonetheless converting written to spoken language. In some ways this fragmentary reading is worse than text reading since the fragments do not have the discourse coherence of complete texts, resulting inter alia in poor intonation. Nor do the coursebooks encourage the students to practice the skill of silent reading.

Though it goes outside the brief here, many of the tasks are essentially those used in the primary school – listing, classifying, mapping, and the like, hardly stretching adult intelligence and interests. The fact that they are carried out in a second language is .arbitrary; there is no logical reason why most of them have to be done through a second language.

The aim here was not to disprove the priority of speech so much as to raise legitimate doubts about the unacknowledged neglect and distortion of writing in the early stages. Language teaching should not be restricted by an invisible straitjacket of assumptions unquestioned since the nineteenth century. If there are indeed powerful reasons for the devaluing of writing, they need to be put forward explicitly.

By neglecting written language, teaching is depriving students of an important element in their lives and minds in a way that other academic subjects do not; music may be pure sound but music teaching in the western tradition emphasises written notes. Putting written language in second place ignores its importance in the minds of literate people. It is short-sighted to make literate students do without its help. Kern (2000) points out that writing develops the ability to think, helps learners to acquire appropriate form/meaning relationships, allows them to develop their ideas, and encourages imaginative use of language. Above all, literate learners use written language as their record of what has happened, to be resorted to when needed, whether the vocabulary list they have made for themselves or the notes they have made in class.

The overall conclusions for language teaching are then twofold. Firstly it should try to curb unnatural practices in the written language. Even if the goal is eventual written language alone, this misuse will have a bad effect on the literate L2 learner. Secondly it should attempt to teach the distinctive elements of writing that are needed by any student who wants to use the second language for more than filling in forms. Written language is at the centre of many uses of language, particularly in the computer age and its characteristics are no more likely to be assimilated unaided than are those of the spoken language.

Many helpful comments and criticisms have been made of earlier drafts by Tony Lilley, Ignazia Posadinu, Guy Cook, Martin Bygate and Brian Tomlinson.

Albrow, K.H. (1972). The English Writing System: notes towards a description. Harlow: Longman

Banathy, B.H. & Sawyer, J.O. (1969). The primacy of speech: an historical sketch. Modern Language Journal, 53, 537-44

Benz, C. & Dworak, M. (2000). Tapestry 1 Listening and Speaking. Boston: Heinle and Heinle

Brookes, A. & Grundy, P. (1998). Beginning to Write. Cambridge: Cambridge University Press

Bygate, M. (1998). Theoretical perspectives on speaking. Annual Review of Applied Linguistics, 18, 20-42

Carney, E. (1994). A Survey of English Spelling. London: Routledge

Chikamatsu, N. (1996). The effects of L1 orthography on L2 word recognition. Studies in Second Language Acquisition, 18, 403-432

Chomsky, C. (1970). Reading, writing and phonology. Harvard Educational Review, 40, 2, 287-309

Chomsky, N. (1995). The Minimalist Program. Cambridge, Mass.: MIT Press

Coleman, A. (1929). The Teaching of Modern Languages in the United States. New York: American and Canadian Committees on Foreign Languages

Courtillon, J. & de Salins, G-D. (1995). Libre Echange. Paris: Hatier/Didier

Cook, V.J. (1989). The relevance of grammar to the applied linguistics of language teaching. Trinity College Dublin Occasional Papers, 22

Cook, V.J. (2001a). Knowledge of Writing. International Review of Applied Linguistics, 39, 1-18, 2001

Cook, V.J. (2001b). Second Language Learning and Language Teaching. London: Edward Arnold. Third edition

Cook, V.J. (2002). Background to the L2 user perspective. In Cook,, (Ed.). Portraits of the L2 User. Clevedon: Multilingual Matters

Cook, V.J. (on-line). The 1400 Sample of L2 Learners' Spelling Mistakes

Cook, V.J. (to appear). The English Writing System. London: Edward Arnold

Cross, D. (1992). A Practical Handbook of Language Teaching. Hemel Hempstead: Prentice-Hall

Cuban Ministry of Education (1999). Principios que rigen la ensen Í anza del ingles en la escuala media. Ministry of Education, Cuba

Derrida, J. (1992). Acts of Literature, (ed. D. Attridge). New York: Routledge

Derwing, B.L. (1992). Orthographic aspects of linguistic competence. In Downing, P., Lima, S.D. & Noonan, M. (Eds.). The Linguistics of Literacy. Amsterdam: Benjamins, 193–210

Goody, J. (2000). The Power of the Written Tradition. Washington, D.C.: Smithsonian Institution Press

Greenall, S. (1994). Reward. Oxford: Heinemann

Guarnuccio, C. & Guarnuccio, E. (1997). Ci Siamo. Victoria: CIS Heinemann

Halliday, M.A.K. (1975), Learning How to Mean, London: Edward Arnold

Halliday, M.A.K. (1985). Spoken and Written Language. Oxford: Oxford University Press

Harmer, J. (1998). How to Teach English. Harlow: Longman

Haynes, M. & Carr, T.H. (1990). Writing system background and second language reading: a component skills analysis of English reading by native-speaking readers of Chinese. In Carr, T.H. & Levy, B.A. (Eds.). Reading and its Development: Component Skills Approaches. San Diego: Academic Press, 375-421

Holm, A. & Dodd, B. (1996). The effect of first written language on the acquisition of English literacy. Cognition, 59, 119-147

Householder, F.W. (1971). Linguistic Speculations. Cambridge: Cambridge University Press

Howatt, A. (1984). A History of English Language Teaching. Oxford: Oxford University Press

Katz, L. & Frost, R. (1992). Reading in different orthographies: the orthographic depth hypothesis. In Frost, R. & Katz, L. (Eds.). Orthography, Phonology, Morphology and Meaning. Amsterdam: Elsevier, 67-84

Kern, R. (2000). Literacy and Language Teaching. Oxford: Oxford University Press

Krashen, S. (1993). The Power of Reading. Englewood Colorado: Libraries Unlimited Inc.

Kress, G. (2000). Early Spelling. London: Routledge

Lado, R. (1964). Language Teaching: A Scientific Approach. New York: McGraw-Hill

Luria, A. (1976). Cognitive Development: Its Cultural and Social Foundations. Cambridge: Cambridge University Press

Lyons, J. (1968). Introduction to Theoretical Linguistics. Cambridge: Cambridge University Press

Nunan, D. (1995). Atlas. Boston: Heinle and Heinle

Nunberg, G. (1990). Linguistics of Punctuation. Stanford: CSLI

Olson, D.R. (1996). Toward a psychology of literacy: on the relations between speech and writing. Cognition, 60, 83-104

Paap, K.R., Noel, R.W. & Johansen, L.S. (1992). Dual-route models of print to sound: red herrings and real horses. In Frost, R. & Katz, L. (Eds.). Orthography, Phonology, Morphology, and Meaning. Amsterdam; Elsevier, 293-318

Petersen, K.M., Reis, A., Askelöf, S., Castro-Caldas, A. & Ingvar, M. (2000). Language processing modulated by literacy: a network analysis of verbal repetition in literate and illiterate subjects. Journal of Cognitive Neuroscience, 12, 3,

Pike-Baky, M. (2000). Tapestry 1 Writing. Boston: Heinle & Heinle

Richards J.C. (1998). Changes. Cambridge: Cambridge University Press

Saenger, P. (1997). Space between Words. Stanford: Stanford University Press

Scrivenor, J. (1994). Learning Teaching. Oxford: Heinemann

Seely, C. & Romijn, E.K. (1998). TPR is more than Commands. 2nd edition, Berkeley CA: Command Perfomance Language Institute

Seidenberg, M.S. (1992). Beyond orthographic depth in reading: equitable division of labour. In Frost, R. & Katz, L. (Eds.). Orthography, Phonology, Morphology and Meaning. Amsterdam; Elsevier, 85-118

Soars L. &. Soars, J. (1993). Headway Elementary. Oxford: Oxford University Press

Stern, H.H. (1983). Fundamental Concepts of Language Teaching, Oxford: Oxford University Press

Sweet, H. (1899). The Practical Study of Language. London: Dent

Taeschner, T. (1991). A Developmental Psycholinguistic Approach to Second Language Teaching. New Jersey: Ablex

Terrell, T.D., Egasse, J. & Andrade, M. (1990). Dos Mundos: a Communicative Approach. New York: McGraw-Hill, 2^nd edition

Tomasello, M. (1999). The Cultural Origins of Human Cognition. Cambridge, Mass.: Harvard University Press

Treiman, R. (1993). Beginning to Spell. Oxford: Oxford University Press

UNESCO Statistic Yearbook (2000). http://unescostat.unesco.org/en/stats/stats0.htm

Ur, P. (1996). A Course in Language Teaching, Cambridge: Cambridge University Press

Valdman, A. (1966), Introduction. In Valdman, A. (Ed.). Trends in Language Teaching. New York: McGraw-Hill