Interview - M A K Halliday
[with G. Kress, R. Hasan and J. R. Martin;
edited by R. Hasan & J. R. Martin;
source: offered by J.R. Martin, 2005-03-29]
GK Well Michael, the first question is why linguistics? We've heard you say that you first turned to linguistics because of frustrations you felt with the way people talked about language in literature classes. What actually frustrated you about literature teaching and why kinds of answers did you find were available in linguistics at that time?
MAKH That was at school where I was trapped in a system which, in a way, I still find unbelievable. It was so over-specialised that from the age of about fourteen I was doing nothing but classics, twenty seven hours a week out of thirty three, and the others were in English. The English part I liked because it was literature and I enjoyed it very much, except when they started telling me something about language in literature. It just made no contact with what was actually there. And this worried me just as it used to worry me when people made folk observations about phonetics; I mean the kind of attitudes Barbara Horvath (1986) observed in her studies of Australian English - for example, that Australians are nasal. It is absolutely wrong, of course, but it takes time to see through these popular beliefs. You asked what was available in linguistics: the answer is - nothing. I didn't find any linguistics, as such; I just went to the library and found a book by someone called
JM What, even in a high school?
MAKH Yes, it was in the library. But I didn't get very far with it. I could write the critical essays when I found out what attitudes you were supposed to have, but I always thought there must be something else - some other way of talking about literature. I felt that there was more to it than what I was hearing.
JM Did you think that language might provide a key or perhaps some kind of objective way of getting access to what literature was about?
MAKH I doubt whether I could have formulated it in those terms, but I felt that literature was made of language so it ought to be possible to talk about that language. After all, my father was enough of a grammarian for me to know there were ways of talking about language. He was also literary scholar although he didn't particularly combine the two in so far as I am aware. I certainly wasn't far enough into it to be able to be more explicit - I think it was more prompted by trying to be more explicit, but as Jim says, I was trying to interpret some of the comments about the language of the work.
RH But when did you make your real contact with linguistics, Michael? When is it that you actually began to feel that linguistics has a possibility for providing answers?
MAKH Well, it was through language teaching. When I left school, it was to take the services' language training course. They took us out of school about eighteen months before we were due for national service, to be trained in languages. I was just seventeen when I left school and joined this program. Now those courses were being run at SOAS. During those eighteen months we certainly heard the name of Firth and we heard that there was such a thing as linguistics. But I don't think I learned anything about it. The initiative had originally come from Firth at the beginning of the war, who said that there was obviously going to be a war in the Far East and in
Now I had in fact wanted to do Chinese anyway and I came out alright on the ones which favoured Chinese so I got my choice. But I presume that if somebody had put Chinese first and it turned out that they couldn't hear a falling tone from a rising tone, they'd have switched them into Persian or some other language.
JM And that was how you really got into Chinese?
JM Before that, you hadn't studied it anywhere?
MAKH No. Apparently for some reason - I have absolutely no idea why - I had always wanted to go to
JM Oh really.
MAHK So I'm told. Apparently I wrote a story when I was about four years old about a little boy who went to
RH Yes, that story is really very, very fascinating. Michael's mother showed me. It has parts of
MAHK I hadn't studied Chinese at all. I really wanted to do Chinese to get out of classics; that was the main motive. I just hated classics at school and I wanted to get out.
GK So you must have been very good at languages to have been called up for this test?
MAHK Well, I don't think you had to be very good. It was just that you had to show that there was some chance you might possibly learn a language. So anyway, they gave us this eighteen months training and we then joined up with the services and I served a year and a half training and then about a year and a half overseas in India. After that year and a half, a small number of us, four out of the whole group that had learned Chinese, were pulled back to
This was 1945 and they thought that there were years of war ahead against the Japanese. And so they increased the numbers of people being trained for the three services. But they needed more teachers; so what they did was to bring back four of us who had done well in the first batch. So John Chinnery, who is now head of the department in
Anyway, even at that time I still wasn't studying linguistics, but I was becoming aware that something like linguistics existed and that there was rather a good department of linguistics just down the street.
GK We've got two questions that follow up your comments - one is about language teaching and how that led into questions about linguistics, and one is about Chinese. First Michael, you've characterised yourself on numerous occasions as essentially an applied linguist who pursued linguistic theory in search of answers to questions posed by language teaching: teaching Chinese to English speakers and later in China teaching Russian and English to Chinese. Initially what was the nature of these questions and the teaching problems that posed them?
MAKH Well, I was brought back to
RH It was more like a realisation "these things that I thought would work didn't work".
MAKH Yes. I had to explain things and I had the advantage of teaching a language which isn't your own and which you've only fairly recently learned; so at least you've formulated some of these problems for yourself - some questions about the structure. I think I began with very straight-forward questions about the grammar because there were so many things in Chinese grammar which just simply weren't described at all and we had been told nothing about them because they just weren't within the scope of traditional grammars and existing grammars of Chinese. We just had to discover them for ourselves. Now I felt very well aware of these and wanted some way of studying them. So this was the first attraction to linguistics, before any other kind - the attraction of educational or pedagogical questions which arose in my mind.
GK Where did you get this consciousness about the problems of Chinese from? Had you been with Chinese people in
MAKH Well, I had just under two years as a student of Chinese. So first of all I got aware of the problems simply as a learner, making mistakes, and asking in the usual way, "why didn't this work?" -making the wrong generalisations the way that a learner does. But then during the time that I'd served in
GK Well perhaps we can go to the next question which is about
MAKH This continues from what we were just talking about. I taught Chinese then for those two years while still in the army. It was particularly during that time that I became interested in Chinese studies generally. Then what happened was that Eve Edwards, who was the Professor of the Chinese department, and Walter Simon, who was then the Reader, felt they had these people who might be interested in studying the language properly. So they organised it in such a way that we taught our courses in the morning and we studied Chinese in the afternoon: all the Chinese courses given by the department for its students were scheduled in the afternoon.
Now at that time, you could specialise in either modern or classical Chinese. I was obviously interested in modern Chinese, so we did a lot of modern Chinese literature and what we could by way of conversation. When I came out of the army in 1947, I decided that I wanted to go on and study the language and the sensible thing to do seemed to go and do it in
JM What year was this?
MAKH 1948. They administered the
So I went around with a young Chinese who was an accountant helping them to keep the books and I wrote publicity. I did this for about six months and then, in some very very small village up in northwest
So anyway the letter said "Proceed back to Peking immediately", and the conditions were that I could spend two more years in
So I re-enrolled at
RH What was his name.
MAKH Luo Zhaunpei, Professor Luo; he died in about 1957. He took me on and I really appreciated this. I wrote essays for him and studied with him. I also went to other seminars.
GK Small tutorials?
MAKH Yes, they were. I can't remember that there was anything like a graduate course; it was more tutorial type of work with groups.
JM Had you done some linguistics back at SOAS?
MAKH No, actually not, none at all.
JM This was the beginning?
MAKH This was absolutely the beginning of it, this study with Luo Zhaunpei.
GK Was there an indigenous Chinese linguistics?
MAKH Yes there was. He knew it very well and it had a very strong tradition going back to the third or fourth centuries BC. However it didn't deal with grammar. Since there's no morphology in Chinese, traditional Chinese linguists never go into grammar. There was a very strong and very abstract phonological tradition which goes back about two thousand years, and as well there was a lexcicographical and encyclopedic tradition. There were these two traditions, yes, but not a grammatical one.
JM What was the linguistic background of your teacher?
MAKH He had been trained in comparative historical linguistics.
MAKH I can't remember exactly. I think it very likely that he would have been in
JM Was there any influence of Sapir and the other American linguistics?
MAKH With Luo Zhaunpei I didn't get into this at all; but it became clear to him after six months or so that I really wanted to work more in modern studies. My own idea had been to work on Chinese dialects. I was very interested in Chinese dialects and was beginning to know something about them. So he said "Well then you need to go and work in synchronic studies; you should go and work with my friend Wang Li". So I said "All right, thank you". I assumed he was across the street, but in fact, he was in
This was in May '49. So I did altogether about seven months with Luo Zhaunpei. You couldn't travel down the country of course because there was very heavy fighting; actually the last big battles were in that very month. So I took a boat out to
Wang Li at that time was the Dean of the Faculty of Arts in Ling Nan, which was a private university. He took me on and that was really where I got into linguistics, through dialect studies. We did basic dialectology, field work methods, and a lot of phonetics, thank goodness. I am deeply indebted to Wang Li for having really made me work at the phonetics and phonology and also sociolinguistics - the whole notion of language in social and cultural context. All those were his contributions.
JM What kind of linguist was he? Did he have a more modern, synchronic background?
MAKH Yes, he had actually been trained in
GK You said just now that the linguistics you studied with Wang Li included sociolinguistics. Can you say something about how he talked about the area of language and social context?
MAKH There was an input from different places by this time. During this period I had become gradually and indirectly aware of some of Firth's notions and while in Canton, I think, I had actually read something of his - what finally came out as his paper 'Personality and Language in Society'. I can't remember how I'd got hold of it. It might have been through Wang Li. Some way or other I'd got some of Firth's ideas and I think Wang Li himself knew some of Firth's work. That was one input. Then secondly, of course, for political reasons, I had become very interested in Russian scholarship. Again, this had started already in
Slavonic linguistics generally has always interested itself in issues such as the development of literary languages and national languages. So that was the second input. So there was the Firthian input and there was that one; and then Wang Li himself as a dialectologist was interested primarily in regional dialects, but was also interested in changing dialect patterns and the social background to these, the spread of the standard language in China, areas of contact between different dialects and the social patterns that went with them. So there were those three parts to it.
GK So although you got your first interest in linguistics in
MAKH Well it was fairly mixed because of all the Chinese linguists, Wang Li was the one who knew most about the Chinese tradition. One of the things that I read and was very much influenced by at the time was his own History of Chinese Phonology, which is a marvellous book. It was so simple, but so very scholarly. He showed how Chinese phonology had developed from the first century of the era through to the tenth century and how it had developed as an indigenous science and then been influenced by the Indian scholarship which came into
JM How long were you in
MAKH A year - well, I arrived in September and left the following May, so essentially a sort of academic year.
JM Was your own research taking some sort of direction at that time?
MAKH Yes, it was actually dialect field work because Wang Li was doing a survey of the dialects of the
I wrote this questionnaire with a large number of sentences and I got them to give me the versions of these sentences in their own local dialects. When I went back to
RH So what happened to all that data?
MAKH It's lying around somewhere; but I couldn't interpret it now I don't think. It's all written in local characters.
JM So you were already a grammarian even by that stage.
MAKH I was really very fascinated by the differences between Mandarin and Cantonese grammar and then how these very local dialects differed in their grammar from the Cantonese. It was very interesting.
GK Do you think your interest in lexis and grammar comes in some way from Chinese traditions in linguistic scholarship?
MAKH I don't think so, because in those days, there wasn't a tradition unless you want to say that I was interested precisely because there wasn't anything there and therefore it had to be filled. But I don't think so. I think I was always basically interested in grammar.
GK What about the lexical part?
MAKH Well, there is one point which hadn't occured to me before. The earliest Chinese work of lexical scholarship is in fact a word list from about 250 BC, which is a thesaurus, and I was always interested in the thesaurus as a tool of lexicography. I have no idea whether there is any connection between those two. It had nineteen different topic headings and lists the difficult words under those headings.
That year with Wang Li was just marvellous. He died recently, just in May - just within the last month. I saw him a couple of years ago - he was a marvellous man and very kind.
Now the terms of the scholarship then were that I went back to
GK Out of where?
MAKH Out of the SOAS, totally - both the Chinese department and the linguistics department.
MAKH Well that's another story. I had left
They asked me when I went for the job at S.O.A.S. whether I was a member of the Communist Party. I said "No", which I wasn't. Then they asked would I undertake that I would not become a member of the Communist Party. I said "No, I wouldn't". So I didn't get the job. When I then asked the person who had questioned me about that afterwards whether that was the reason, he said "Political considerations were not absent". I thought this was the classic answer of all time. So the point is that I got shunted off to
MAKH The Chinese department at
JM How disappointing was this for you? You had particularly wanted to study with Firth.
MAKH It was very disappointing because I wanted to study with Firth and I wanted to work on my dialect material. The price of going to
My supervisor was Hallam but I was negotiating with him to be allowed to go up to
JM Devious ways he finds to ride on trains!
MAKH Yes, that's when I started finding you could work on trains.
RH But this is extraordinary. They didn't allow you to stay at SOAS because you wouldn't give an undertaking not to enlist in the Communist Party; and then you still came back, and you were still on the premises.
MAKH Yes, but I wasn't teaching. That's what they were scared of! I was not in a position to subvert.
GK So your first contact with Firth had been fortuitous but when you returned from
RH That's asking the entire history.
MAKH "Interested in his framework". Okay. From the start when I became gradually aware of his ideas, particularly I think during that year with Wang Li, I felt very sympathetic. It seemed to me that he was saying things about language that made sense in terms of my own experience and my own interests, and I just wanted to explore those ideas further. My main concern was just to learn from him and I managed to organise it so that he took me on officially as a student. What I got of course from him was enormous, both in terms of general philosophical background and insight into language. But I didn't get a model of grammar because, as you know, Firth himself was interested in the phonology, semantics and context. He had very little to say about grammar, although he certainly considered his basic system/ structure approach was as valid in grammar as it was in phonology. My problem then, as it seemed to me, was how to develop system/structure theory so that it became a way of talking about the language of the Secret History. Now the text was a corpus - for Firth it was a text and that was fine. That meant it had its own history and had to be contextualised and recontextualised and so forth. It was also closed, in the sense that you couldn't go out and get anymore. This was 14th century Mandarin and that was it. There wasn't any more. So you treated it as it was. I was not yet, of course, aware enough to be able to ask questions about what it meant to consider it just as a text as distinct from considering it as a instance of some underlying system. But I tried to work out the notions of system and structure on the basis of what I read and what I got from Firth in phonology.
JM Was W. S. Allen working on applying Firth's ideas to grammar in this period too?
MAKH Yes, although I didn't actually get to know him very well. The person who helped me most other than Firth at that stage was Robins. In fact Firth got Robins to do some of the supervisions for him. I used to write essays for Robins and so forth. Robins was terribly nice and very helpful. But I didn't know Allen very well.
JM Robins was on staff there?
JM Allen was also?
MAKH Allen was, yes. All that generation was there. Of course some were still students.
JM When did you have a chance to see 'System and Structure in the Abaza verbal complex'?
MAKH That was not until after I finished my thesis.
JM So you really had to do this all on your own.?
RH When did you finish the thesis?
MAKH At four o'clock on the last day after the last extension, I can tell you that. It was an hour before they closed the offices and it was the 31st December. I can remember that, and it would have been in 1954.
JM So you spent three years in
MAKH Four years, because it was 1950 when I moved to
JM And Robins, had he been thinking of applying system/structure theory to grammar?
MAKH Not, I don't think so. He was more interested in phonological applications. I was very much on my own at that. It wasn't that there wasn't any place for the grammar for Firth. He would introduce examples in his lectures - for instance working through the forms of the German definite article as a way into raising a whole lot of interesting grammatical problems. And he was developing, at that time, the notion of colligation, which actually Harry Simon labelled for him. But it was never very much developed so that its not terribly clear what Firth was ever planning to do with it; but it was the beginnings of thinking about grammar.
GK Who were the other students at that time?
MAKH I'm not sure which year different ones were there but certainly listening to Firth lectures at different times during this period were, for example, Frank Palmer and Bill Jones who were themselves just getting onto the staff of that department, Bursill-Hall who then went to Canada. Mitchell was already on the staff, as were Robins, Allen, Cornochan and Eileen Whiteley. I also went to other lectures when I could - to Eileen Whitely for example and to Eugenie Henderson. I got a lot of phonetics from them as well as other things. Eileen Whitely never wrote anything, but she just had a fantastically broad range of interests. She was one of the people who really could have developed Firth's notions, especially in the direction of text - in a sematic direction. She was very very good.
GK Could we return to that question about how you went about extending Firth's ideas so that they could be applied to the grammar of a language?
MAKH I tried to understand, not always very successfully, the key notions that could be interpreted at a general enough level so that they could be applied to grammar as well as to phonology. For example the concept of prosody - the notion that syntagmatic relations pervaded items of differing extent. Firth as you know was concerned that you didn't start and end with phonemes, and so forth. Rather you looked over the larger patterns and then residual effects, so to speak, were handled at the smaller segments. Now I tried to apply that idea to grammar, so I began at the top. That's one very clear example, using a kind of topdown approach, beginning with what I could recognise as the highest clearly defined unit in the text, a clause, and then gradually working down. Then another basic concept, of course, was the system/structure notion, which I found very difficult - especially expressions like systems giving value to the elements of structure.
I tried to set up structures in a framework that was formal in the sense that you were not relying on some kind of vague conceptual label. For example, it was possible in Chinese grammar to set up categories of noun and verb on distributional grounds. These then gave you a basis for labelling elements in the structure of the clause.
RH How important was the idea of exponence for Firth?
MAKH Well, it was very important. You see, there are a number of ways in which I built on his ideas that he certainly wouldn't have followed, as he made clear to me. I got on well with him and he didn't like people who weren't prepared to argue with him. But of course the cost of this was that I would often be seizing on things that he'd done and, from his point of view, misinterpreting them in some ways, in order to try and do something with them. Exponence would be one example of this. Firth had a long running argument with Allen in 1954-1956 about the nature of exponence and about the relation between the levels and exponence. As far as Firth was concerned the levels (the phonetics, the phonology, the morphology, and the grammar or whatever) were not stratified but were rather side by side, each directly relatable to its exponence. So you didn't go through phonology, so to speak, to get to the grammar. On this point Allen disagreed about the nature of this pattern. As far as Firth was concerned, there was absolutely nothing wrong with using the same bit of datum over again in setting up patterns of different strata, whereas Allen seemed to say "Well, if you'd built this particular feature in to your phonetic interpretation, you couldn't use it again in the phonology". So there were differences of this kind in the way they worked out this notion, and Allen's, in fact, was the more stratified view, although I don't think he expressed it like that. I did not follow Firth on that because I just couldn't see any way that you could get the notion of realisation into the grammar except by stratifying (although I didn't use the term realisation then).
So exponence for me became this kind of chained relationship, which it was not for Firth.
2. Grammatical theory
GK We would like to ask you about the grammar and our first question is about the focus on system. We think you are a great relativist and unusually modest about the claims you make for systemic theory. Your theory gives greater prominence to paradigmatic relations than any other. What are the strengths and weaknesses of this focus?
MAKH Well, I didn't start out that way of course. Because that links back to what we were saying about Firth. As you know, for Firth, there was no priority between system and structure - they were mutually defining. Indeed, if anything, in the context of linguistics of the time, his emphasis was on the importance of syntagmatic relations.
So in a sense, I'm going against Firth. Now why was this? Firth himself didn't really believe in "The System" in the large sense at all. His interest was not in the potential but in the typical actual. Now this meant that for him, in fact, the priority was to structure over system - not in the structuralist sense of language as an inventory of structures, but in the sense that, as he put it, the system is defined by its environment and its environment is essentially structural.
So in a sense, the larger environment is the syntagmatic one. Now trying to work this out in Chinese grammar generally, I felt that I needed to be able to create the environment that was needed. The environment had to be set up in order for the general framework of the grammatical categories to make sense. But this environment seemed to me ultimately to have to be a paradigmatic one. That took a lot of steps - say 1962 when I was writing 'Syntax and the Consumer', or 1963, when I was doing the 'Laundry Card Grammars' in
By various steps, I came to feel that the only way to do this was to represent the whole thing as potential - as a set of options. And this was certainly influenced by my own gut feeling of what I call 'language as a resource' - in other words, language was a mode of life, if you like, which could be represented best as a huge network of options. So that kind of came together with the notion that it had to be the system rather than the structure that was given priority.
GK How do you see that now?
MAKH Well, in an important sense I would think that there are a lot of purposes now for which it's important. Just to mention one of them, I think that in order to crack the code, as a probabilistic system, we have to start with a paradiagmatic model. It doesn't make sense otherwise. But, of course, it does beg a number of questions in a sense - this is something we often talk about. The great problem with the system is that it is a very static form of representation. It freezes the whole thing, and then you have to introduce the dynamic in the form of paths through the system. Your problem then is to show how the actual process of making paths through the system changes the system.
This is crucial to the understanding of ontogenesis, phylogenesis - any kind of history. So I think I shall continue within that framework because that's the one I'm familiar with and I've not enough time to start re-thinking it.
GK In the era of post-structuralism Firth seems more contemporary than you. I mean I already have problems with post- structuralism and the dissolution of system, but that is the contemporary flavour of thinking about things.
MAKH I often get the feeling that all these -isms, whereever they raise their head, want to go too far either in this direction or in the other direction. In practice it is just not possible to have systems without the product of those systems, which are the structures; which means that the structures must be there do deduce the system. How far do we go back in this kind of argument? Either you're forced to the point where you say the entire system is, was, has always been, or you have to say that in some sense, structure, which is a constraining name for process, is where it all begins. Because otherwise you can't have systems.
I would comment that these things obvious switch between extremes. There is an important sense in which you can deconstruct the system, as it were; you can remove it from your bag of tools. But you have to get it back again, if only because you can't deconstruct something if it isn't there; so there's no meaning in doing so. I think we are now at a stage where we are realising that the models we have to look at for systems are not solely in the areas this kind of post-structuralist thinking is reacting to. Their critique has almost become irrelevant, I think, in the light of much more general developments in modern scientific thinking, which really transcend the differences between human and non-human systems. Once you start looking at systems in this sense, you have to have the concepts. Take for example Jay Lemke's work in dynamic open systems. This is the sort of thing that I find interesting as a way of looking at language. And the sort of work that's being done in physics as well is totally annihilating the difference between human systems and sub-atomic systems.
GK We wanted to ask you about strengths and weaknesses. Do you see any weaknesses in that greater focus on system rather than structure?
MAKH Well, one I've mentioned is that it's overly synoptic. I mean it's static. Also there is the danger of its pushing the system too far apart from process/text. I mean I've tried to avoid doing that. It's one of the things that Firth so strongly objected to in Saussure - the dichotomy of langue and parole which prevents you from seeing that langue is simply a construct of parole. I would agree with that and I think that there is a danger of using system as a tool for thinking with and forcing a kind of dichotomy between the system and the text. I think those two are dangers really.
GK We've got a question about function: since the late sixties, systemic grammar has always been for you "systemic functional grammar". What is the relationship between different concepts of function (for instance 'grammatical function', 'meta-functional component, and the natural relation you propose between metafunction and register) that you use? And just how critical is their place in your model?
MAKH I think they're important and I think they're closely related. I have usually felt that the best way of demonstrating this relationship is developmentally because you can actually see, if you follow through the development of a mother tongue, how the system evolves in functional terms. In the beginning, function equals use, so that there is a little set of signs which relate to a simple theory, on the part of the infant, that semiosis does certain things in life. You can then watch language evolving in the child in this context.
So the metafunctions are in my view, simply, the initial functions which have been reinterpreted through two steps. The first involves generalisation: initial functions become macro-functions, which are groupings which determine structure. Then macro-functions become metafunctions: modes whereby the linguistic semiotic interfaces with contexts. So I see this as very much homogeneous. The notion of the context plane as something natural is part of the same picture.
GK Can you just expand on that last phrase?
MAKH If language is evolved as a way of constructing reality - then it is to be predicted that the forms of organisation of language will in themselves carry a model of that reality. This means that as well as being a tool, language will also be metaphor for reality. In other words, the patterns of language will themselves carry this image, if you like. This is what I would understand by talking about a 'natural' grammar.
RH Would you sat that's another way of saying that reality is the product of semiosis.
RH And in that sense the question of a 'natural' relation between the grammar and the reality that it constructs has to be either answered 'yes' or it becomes a meaningless question?
MAKH Okay. Right. I mean that reality has to be constructed, so it's another level of semiosis. So it's inevitable, in a sense, that the semiotic that you use to construct it will, in some sense, replicate that which you are constructing with it, since it's all part of the same process. I want to be rather cautious on this. I think we're in a phase at the moment where we are emphasizing this point. We've gone against naive realisms which assumes that there is something out there that is given and that what we have to do is to mirror it in some sense, which is certainly where I started from. And we've kind of moved into a phase of thinking again, the opposite extreme so to speak. We are now emphasising, as you were saying, the fact that it all has to be constructed. It is, in fact, a many levelled semiotic process. And that, in a sense, is an important corrective to naive realism.
JM I was interested in the grammatical functions themselves, Subject, Theme and so on. You use functions and class labels in your model. How crucial is that to this functionalism idea, and this idea of a natural grammar.
MAKH It's part of the picture. In order for the system to work with some kind of output, in other words to end up as speech sounds, signals or writing on whatever, there has to be this re-coding involved in it. The fact that there has to be this re-coding means that there must be a non-identity between functions and classes; otherwise you wouldn't need to re-code: you could do the lot at one level. So somewhere or another you've got to be able to talk about this. Now it seems to me then you have to decide in finding ways of interpreting language how you're going to do this. An obvious example would be formal systems. If the main priority is representing language as a formal system, then presumably you'll prefer a kind of labelling in which you have class labels and conventions for deriving functions from them. For my own part, I prefer theories to be extravagant and labelling systems to be extravagant. As a tool for thinking with, I've always found it useful to separate function and class and build that amount of redundancy into the discourse about language. It then becomes possible to operate with sets of functional labels in the grammar, things like Theme and Subject and so forth which, in turn, enable you to make the links outside. So I think it is a useful and important part of the whole process. There is a reason for wanting to separate these two, although if you focused on any one specific goal, as distinct from trying to keep it all into focus at the same time, you could do without it. And I think I would say this as a general truth. There's very little in what I've done, or what is done in systemic theory if you like, that couldn't be done more economically in some other way if that was all you were interested in doing; and, I suppose, what I've always been concerned with is to work on little bits in a way which I then don't have do abandon and re-work when I want to build them into some general picture.
GK I think this relates a bit to what we were saying earlier where you were talking about system and structure. The question is: In your model the relation between various components, between strata, between ranks, between function and class, between grammar and lexis, is handled through the concept of realisation. This involves, in English at least, setting up a Token/Value structure with the component closer to expression substance as Token and the component closer to content substance as Value. This gives the Value component a meaning of temporal priority, apparent agentivity, greater abstraction, greater depth and so forth. Is what English does to this concept, in fact, what you mean by it?
RH In raising this question we were trying to build in the informal kind of discussions we've been having recently on realisation. You've argued very strongly that when we say 'x realises y' then, in some sense, because of the structure or whatever, you get a pre-existence postulate there which you would like to deny. This seems to me a very important point. To my way of thinking it also links ultimately to system and structure, to the langue and parole question, and is altogether the most central concept in the whole theory.
MAKH I'm with you. It is absolutely fundamental. Maybe we could have a workshop, an International Systemic Workshop, just on realisation. That would be nice.
You know, the problem is you can't talk about it in English. Not only the temporal priority but of course the agentive priority gets in the way. I mean the Agent is the Token. According to the grammar of English it's the Token that does the work so to speak. I started with a fairly simple notion of something 'out there' to be realised through the code. It's, again I think, something that we have to think of in the light of recent thinking about the universe we live in as an information system. And what English does to the concept, I think, is a very important clue. I mean, what any language does to the concept has to be taken as a very important clue, a way of thinking about it. And again it's at this point that the grammar as a tool for thinking about other things becomes crucial. I think linguistics has got to accept its responsibility now, as being the core science. In a sense it has to replace physics as the leading science.
RH There is another problem here. If you think in terms of languages that in their structure are very very different from Indo-Eurpean languages, well then you might expand this discussion. So in some sense to me the problem becomes circular. We perceive that there is this problem for expressing the relation of realisation of structure of English, and yet we cannot yet bring evidence from any other language that it could be otherwise, because by our way of talking, we will impose a pattern on that language.
MAKH In a sense it's one of these things that probably has to be done before it's too late. What happens in practice is that people tend to borrow English (or whatever the international language is) ways of talking about things, and you want to know how they would develop otherwise.
GK We have a question which is around that point of grammar and linguistics and the language shaping both the linguistics and the theory - what you think of as grammar symbolising reality. Following from the point about realisation that we made in our earlier question, to what extent have the meanings available in English or Chinese consciously or unconsciously shaped your model of language?
MAKH Let me answer that quite quickly. I'm sure they have and I've tried to make it conscious. It's impossible that they couldn't so I have tried to be conscious that they are shaping it when thinking about it. One of the things I regret most is never having been able to learn another totally different language. I made two attempts to come to
JM Chinese and English weren't different enough?
MAKH Not really. They are different in interesting and important ways but they both have a long written tradition.
GK In terms of that question about realisation, it would be nice to have a language that was far more verbal and not written, to understand how people might think about that.
RH Yes. I think if one did this kind of study one would find that writing does another thing - it objectifies in a way that the oral tradition doesn't so that what you would get would be more like myths as metaphor for certain sorts of beliefs, certain sorts of perceptions, instead of this explicit analysis where the concepts are defined, placed in relation to each other clearly, and then you go and talk about their interrelations. That's the way it happens in languages that don't have a literary tradition.
MAKH That's also why we're still stuck with Whorf. I don't mean by that that I want to give up Whorf, as you know; but what I mean by that is we've got nothing else yet. It's easy enough to get the mythologies, the things at that semiotic level, if you like. Now as you move into the grammar what happens is that nearly everyone working on the grammar in these languages is a universalist. So of course, they're interested in making them all look alike; and so you're left with Whorf. And it's in the grammar, you see, that I want to find new models.
JM What about, say, between English and Chinese. I mean, can you point to the parts of your model where it would have been different if you hadn't known Chinese?
MAKH It's very hard to say of course. I suppose that one of the things that is absolutely critical has been that for me grammar has always been syntax, since Chinese has no morphology. I cracked the Chinese code first. There are other things, yes, for example temporal categories.
GK What about that point you made earlier about prosody?
MAKH Yes that, of course, could have come from Firthian phonology without necessarily going through Chinese, although the Chinese helped. But it was Chinese phonology at the time of course.
GK And tone?
MAKH That's true. That's certainly true. Then there's the point of syntax. Then I think there are certain special features about Chinese grammar which did affect my thinking. There was something that Jeff Ellis and I wrote many years ago, which I must see if I still have, because it wasn't a bad article. It was on temporal categories in the modern Chinese verb. It was important because, you see, it was a non-tense language. Jeff was extremely well informed about aspect as he had started off in Slavonic and he had studied aspect systems round the world. So Chinese helped me to think about time relations in a non-tense sort of a way - the Chinese system of phases his a clear grammatical distinction between a kind of conative and the successive; the verb essentially doesn't mean you do something so much as you try to do it. It does not necessarily imply that you succeed. Now I don't read a naive cultural interpretation into that but it forces you to think differently about the grammar.
JM Would the lack of morphology in Chinese, have been something that pushed into paradigmatics?
MAKH Okay, yes. That's a good point. I mean your paradigms have to be syntactic. You can't start with a word and paradigm approach. There are no paradigms and one of my main strategies in working on Chinese was setting up syntactic paradigms. They were there already in that article in 1956 on Chinese grammar.
GK A question about choice. Although you model language in terms of choice, in many respects this choice is almost never free. What is the place of your position on the probablistic nature of linguistic systems in modelling these constraints?
MAKH I have always thought of language, the language system, as essentially probablistic. You have no idea how that has been characterised as absolutely absurd, and publicily ridiculed by Chomsky in that famous lecture in 1964. In any case, one point at a very simple level is that nobody is ever upset by being told that they are going to use the word the more often than they're going to use the word noun. But they get terribly upset by being told that they're going to use the active more often than the passive. Now why is that? We know of course that we have well developed intuitions about the frequencies of a word - and can bring those to the surface. But we can't bring them to the surface about grammar; and in fact all that is doing is just showing that, as always, the grammatical end of the lexico-grammar continium is very much deeper in the (gut) and it's much more threatening to have to bring it out. But it's there. The question then of what this actually means in terms of the nature of the system is extraordinarily complex and it's really does need a lot of thinking and writing up, exploring what it means in terms of real understanding of the nature of probability and statistical systems and so on. Again, what I want to do is try to bring probability into the context of a general conception of systems, dynamic open systems, of what this means in terms of physical systems. It has to be seen in that light as I was saying before.
JM This seems to be something quite critical in your theory, this idea of probablistic systems, especially in terms of not losing sight of the text and the way in which the text feeds back into the system. You have to view text as passes through the system which are facilitating.
MAKH I would agree with this and you have to have this notion in order to show how the system shapes the text anyway. The pass through the system in fact changes the system just as every morning if you turn on your radio they will tell you that the temperature is ten degrees and that's one below average so to speak; but that has changed the average. So everytime you talk, everytime you produce a text, you have of course changed the system.
3. Language in Education
GK The next section is on language development and education. Our first question is about language in education. You've been the driving force behind language us education movements in both
MAKH Well, I come from a family of teachers so I suppose that the whole educational process has always been of interest; and I had my own time as a language teacher. I've always been, if you like, motivated in working on language by the conviction that this had some practical value, and that education was the most accessible in a sense. There are a lot of other applications. Obviously an important one is clinical. But I don't know anything about that, and in any case we were a long way from actually getting linguists working together with pathologists. But it seemed in the late fifties that people were ready to think about language in education.
My first position in linguistics was in the English department at
Each of us had different groups of teachers that we used to work with. Now this was when I came in to mother-tongue education, because these were English teachers in the Scottish schools. It kind of reinforced my feeling that we really needed an input from linguistics.
Then when I moved to London in 1963, the first thing I did was to set up this project with Nuffield money, which became the School Council Project in Linguistics in English Teaching, which was producing Breakthrough to Literacy.
RH But behind this, at a deeper level, didn't you have a feeling that linguistics is a mode of action, that linguistics is for doing rather than just intellectualising?
MAKH Yes, very much. I don't really separate the two in any sense at all. I've always seen it like this. My problem has always been that teachers want results too quickly. In fact the reason why we have to work in this field as academics is that we have a longer term perspective. We can say: "You've got to go back and do so much more fundamental work. You've got to back off for a bit. You can't expect result by next Tuesday". And that's where linguistics comes in. It's a mode of action but it's a mode of longer term action, if you like. You have to have the luxury of being able to look further into the future.
GK We have a question about applied linguistics as a mode of action. Our question asks whether this is an expression of your political beliefs. We have a little aside here which asks whether, like Chomsky, you see linguistics and politics as unrelated spheres and, if not, how it is that you are able to make as much use as you have of Firth's ideas when your politics and his were far apart.
MAKH That is an absolutely fascinating question. You'll have to stop me because I'd love to go for two hours about that. No, I do not see my linguistics and politics as separate. I see them as very closely related. To me it's very much been part of this backing-off movement. In other words, I started off when I got back to
There was a lot of things going on at the time. There was the Menist school; there was the Pravda bust-up in 1950; there were current developments in English Marxism and things of this kind. Later on came the New Left, of course. But it seemed to me that any attempts to think politically about language and develop a Marxist linguistics were far too crude. They involved almost closing your eyes to what you actually knew about language in order to say things. My feeling was we should not. Of course the cost of doing this is that you may have to cease to be a Marxist, at least in a sense in which anyone would recognise you as one, in order to go away for fifty years and to really do some work and do some thinking. But you're not really abandoning the political perspective. You're simply saying that in order to think politically about something as complicated as language, you've got to take a lot longer. You've got to do a lot of work. And you've got to run the risk of forgetting that what you are doing is political. Because if you force that too much to the forefront your work will always remain at the surface; it will always be something for which you expect to have an immediate application in terms of struggle. You can't do that in the long run. You're going to pay the price that you may achieve something that's going to be useful for two weeks or two years and then it'll be forgotten.
I always wanted to see what I was going towards as, in the long run, a Marxist linguistics - towards working on language in the political context. But I felt that, in order to do that, you really had to back off and go far more deeply into the nature of language.
JM You were ready then for teachers' reactions to your ideas? It's the same problem of distancing.
MAKH Yes, it is. Now with Firth, you see, it is very interesting because Firth was right at the other end of the political spectrum. There was in fact another interesting occasion when I went to be interviewed by him for a job at SOAS (not the same as the first one, different in a very interesting way, although with exactly the same result).
It was after this interview in fact that Firth said: "Of course you'd label me a bourgeois linguist". And I said: "I think you're a Marxist", and he laughed at me. It seemed to me that, in fact, the ways in which Firth was looking at language, putting it in its social context, were in no way in conflict with what seemed to me to be a political approach. So that it seemed to me that in taking what I did from Firth, I was not separating the linguistic from the political. It seemed to me rather that most of his thinking was such that I could see it perfectly compatable with, indeed a rather necessary step towards what I understand as a Marxist linguistics.
GK So Firth must have been, at some level, confused - to have contradictions in...
MAKH Does that necessarily follow?
RH: I don't think people's ideologies are coherent.
MAKH: No, that's certainly true.
RH I don't think they are. I think Firth had this ideology about language, it's role in society, about its role in forming people and all that. On the other hand he also had this very strong authoritarian attitude towards institutions and their maintenance and things.
GK A question which relates to all of that - theory out of practice. To what extent has your commitment to applied linguistics influenced your model? And how has it influenced the research that you've pursued?
MAKH Well it's influenced it, of course, in one sense by making sure that I never had time to do much thinking about. Yes, I think it's influenced it. It's hard to say exactly how. I mean, I've always consciously tried to feed back into thinking about language what came from, say, the experience of working with teachers. The Breakthrough materials would be one case in point. I have always tried consciously to build teachers' resources into my own thinking about language; David McKay for example, made an input with observations on children's language learning in an educational context. Then, of course, through Basil Bernstein's research and Ruqaiya's part in his unit, there was another source of input from what, in the broadest sense, is a kind of applied linguistics.
RH Can I stray from the point here? It seems to me that talking to the teachers and the need of making your linguistics accessible to the pedagogical circle had a different influence on your work from that which say, for instance, contact with Bernstein's unit might have had. The first one forced you to write in a way that would make your material accessible. In other words, I do not see that the shape of the theory, the categories as such, got terribly shaped by that, (though it is always a bit doubtful to make these kind divisions). On the other hand I feel that contact with Bernstein's work had an effect of a slightly different kind in that it really fed right into theoretical thinking.
MAKH That is definitely true. I wasn't prepared to shift because of teachers in what seemed to me to be short term directions just because it seemed to be something that would be a payoff in class and so on. So it was more in the form of presentation. But I think there was some input from educational applications.
GK Most of your work has been in mother-tongue teaching and we were wondering how much of this was historical accident, how much by design?
MAKH My first publications relevant to language teaching had a strong E.S.L. focus. In
GK That's true now isn't it, in lots of ways?
MAKH It's true in lots of ways although there are some ways now in which I think mother-tongue teaching is taking over.
GK You've been centrally involved in two major mother-tongue research programs, the Nuffield Schools Council and the Language Development Project work in
MAKH I suppose what has been achieved is a number of fads and fashions, some of which will remain. In the English Language Teaching context it seems to me there are two developments which were applied linguistic developments which were important. One is the notion of language for specific purposes which came quite squarely out of Firth's restricted languages and concept of register. And so I think that's been an important part. I think in the mother-tongue area, two things have been important. One is the awareness of the child as a human being who has been learning language essentially from birth so that the learning in school becomes continuous from that. And related to that perhaps, the notion of language as a process in education. Things have changed. The very concept of language education didn't exist twenty-five years ago, or even fifteen years ago. So I think that most of the achievements have been based on gradually raising the level of awareness of language among educators. One has to remember this sort of thing has to go on, over and over and over again. It doesn't suddenly happen.
GK But are there things now that you don't any longer have to say very strongly that you might have had to say twenty years ago?
MAKH Well, there are some I think, but not very many. I think you have to go on and on and on saying them every year, to each new group of students. I suppose we don't any longer have to fight the old fashioned views of correctness and language as table-manners (again we can't be complacent about these things). And we don't have to fight the notion of standard and dialect, and dialects as being inferior. I think people have moved quite a lot on that. There's a more complicated history as far as relations between spoken and written language are concerned. At one time I would have said we no longer have to fight the battle for recognition of spoken language in education. But I'm not sure about that now. I think we're going to have to gear up for a new battle there, though on a different plane certainly.
Even where one doesn't feel there has been much progress, the discussion may have moved onto a different plane. I think we've always been aware, and it's certainly true now in
GK We've got a question which follows that up a little bit. Your theory has been designed to solve problems or at least to play an active apart in solving them. Which parts of it do you think have been most effective and what are you most proud of in terms of what has been achieved?
MAKH I suppose it ties up with this section generally. I feel that it's been in the educational area. I think I'm a little bit proud of that, and have that feeling on various levels. For example, I first started intervening myself when David Abercrombie said to me, "Will you teach on my summer school, the British Council Summer School for the Phonetics of English for Foreign Students" This was in 1959. And I said: "Certainly. What do you want me to teach?" He said "Well, you know Chinese. Teach intonation". I knew nothing about English intonation, so I started studying it, trying to describe English in such a way that the description was useful to those who were going to be working on language in the classroom, in an educational context. The fact that we are now getting to the point when people are saying: "I can use this grammar for working on language in the classroom." is an achievement. When I went to the Nuffield Foundation in 1963 I said: "I want some money for working on language in this sense, but I don't want to see any teachers for years because we're not ready for them, so to speak". And they said, "If you put the teachers in right away, we'll double your money". You can't refuse that kind of thing. Of course they were quite right. What this meant was that we used to have those weekly seminars, when we had David McKay and Peter Doughty working on grammar, from the point of view of where it was going to be used. Now at that time you didn't dare put it into the program because, certainly in
One of the things I feel most happy about is the developmental interpretation that I tried to put on early language development and the importance of that for later work on language in education. That again, came out of teachers. When Language in Use was taken up in the 'Approved Schools', that is the schools run by the Home Office for children who had been before the courts, the teachers came to Geoffrey Thornton and Peter Doughty and said: "We want to use these materials. Would you lay one a workshop for us". And they asked me to go and talk. At the same time David Mckay and Brian Thompson, who were the Breakthrough team, set up a workshop for primary school teachers. Both groups asked me the same question which was: "Tell us something about the language experience of children before they come into school at all". I hadn't done anything of course at that time but I read around on what there was. Not much of it was terribly useful. Ruth Weir's was one of the best in those days. But it started me thinking on early language development. That was the time when Neil was born and when the Canadian Government wouldn't let us into
RH Those were difficult, perhaps fortuitously difficult times in more than one respect. In some sense the rise of Chomsky's linguistics must have impinged on your work in the sixties and the early seventies. Why did you hold back, in spite of general acceptance of the TG framework, and how did you see yourself in relation to that whole movement?
MAKH Chomsky's work quickly became a new establishment, and in many ways a rather brutal establishment actually. At
The way the goals of linguistics were defined at the time, the notions embodied in all the slogans that were around, 'competence' and 'performance' and things of this kind, I just found quite unacceptable. Intellectually I thought they were simply misguided and in practical terms I thought they were no use. So that I thought that if one is really interested in developing a linguistics that has social and educational and other relevance that wasn't going to help. We just had to keep going and hope that it would wash over and we should be able to get people listening again to the kind of linguistics I thought was relevant.
RH And it happened.
MAKH Yes, it happened, and now we know it'll all disppear into the history of the subject eventually.
4. Language and context
GK We've got a set of questions on language, linguistics and context. Our first question is on politics: You are someone whose career has been disrupted more than once because of your political beliefs. Have these experiences affected your approach to linguistics, especially linguistics as doing?
MAKH No, I don't think so. I mean, yes, okay, I was witch-hunted out of a couple of jobs for political reasons. And the British Council refused to send me any where at all during that time, however much people asked. But I don't think that this has affected my approach to linguistics. Linguistics as doing is part of a political approach and I didn't suffer in the way that a lot of people suffer. Of course I've no doubt that I would have gone in very different directions had this not happened. For example, if I had been taken on and kept on in the Chinese department at SOAS I might well have stayed principally in Chinese studies and worked on Chinese rather than moving into linguistics generally. And secondly, of course, the thing that I really wanted was the job on Chinese linguistics in Firth's department. It was for purely political reasons that I didn't get that. I wish that I had that interview on tape because it would be one of the most marvellous documents ever. It would be fantastic, absolutely fantastic.
RH For the analysis of ideology!
MAKH Yes, it really would be. It was absolutely incredible. In any case if I'd got that, I think, I would have remained much more closely a Firthian. I wanted to get into Firth's department. If I had got into Firth's department, I would have quite definitely have worked much more within Firthian framework. You have to remember that to the extent that I have departed from Firth, certainly initially it was simply because I wasn't there in the group in any sense and therefore I wasn't able to get answers to questions, and, in some ways, to correct misunderstandings,
This meant that, in a sense, I was pushed out to working on my own in two instances where in either case, if this hadn't happened, I might well have continued to work in the pre-existing frameworks, both institutional and intellectual frameworks. I'm not sorry.
GK Our second question is about language and social reality. You are one of the few linguists who have followed Whorf in arguing that language realises, and is at the same time a metaphor for reality. How Whorfian is your conception of language and what part has Bernstein's theory played in shaping your views?
MAKH Well, I think it's Whorfian. Partly this is because you can make Whorf mean anything you want. When I say I think my conception of language is Whorfian, you know what I mean; but for a lot of people who would interpret Whorf differently that might not be the case. I certainly follow some aspects of Whorf's work, which I think are absolutely fundamental. One is the relation of language to habitual thought and behaviour. Another one, perhaps less taken up, but which I think is fundamental, is the notion of the 'cryptotype' where it seems to me that Whorf (and of course in this he was simply following the Boas-Sapir tradition) was so right in the seeing the action at the most unconscious levels. The whole point is that the Whorfian effect takes place precisely because of what is going on at the most unconscious level. And, one might add to that, it's going on in what is an evolved human system and not, as sometimes represented, an artificial system. Language is a natural system. In fact, it is these two things, the naturalness and the unconsciousness which make these effects possible.
I was arguing this with an economist about two years ago. He was saying in effect that it is only through the most conscious forms of human activity that ideologies are transmitted and that social structure and social system is maintained. And he was therefore defending sociological and economic models of research. In other words you go and study how people plan their budgets or do their shopping or whatever. And I was arguing the opposite case. He was saying: "How can you claim that language can have any influence on this because it's all so unconscious". He wasn't disputing that the processes were unconscious but saying that because they were unconscious they could have no effect on ideology. And my view is exactly the opposite - that it's at the most unconscious level that we essentially construct reality. And that, I think, is Whorfian. Therefore, particularly in terms of the grammar, it's the notion of the cryptotype that I would see as absolutely essential.
JM I wondered if Chinese comes in here again in the sense that a grammar of Chinese could only be a grammar of covert categories, because there are no overt ones.
MAKH It never occured to me but it may well be true. I've never thought of that. Now as far as Bernstein is concerned, he himself, as he often acknowledged, also took a great deal from Whorf. He makes the entirely valid point that Whorf is leaving out the component of the social structure. Whorf essentially went straight at the ideational level, from the language into the culture, so to speak. Bernstein has pointed out that there has to be, at least in any general theory of cultural transmission, the intermediary of the social structure. I think this is actually right. Bernstein is still, uniquely as a sociologist, someone who has built language in as an essential component to his theory, both as a theory of cultural transmissions and as a general sociological and a deep philosophical theory. He convinced me that this was possible. Perhaps this hasn't come out clearly from what went before because we talked more about the applied context, educational and other applications. But I think it's important also to say that a representation of language has to be able to interpret language in the context of more general theories of social structure, social processes and so on, and ultimately of the whole environment that we live in. In general that had never been done. In fact, the problem has always been in linguistics that linguists have always shouted loudly for the autonomy of the subject, and that always seems to me to be of very little interest. Linguistics is interesting because it's not autonomous. It has to be part of something else. Now Bernstein was the first person that made it part of something else and so the way in which he did this was obviously important.
I used to argue with Bernstein when he was doing it the wrong way. Early on he was looking for syntactic interpretations of elaborated and restricted codes; I always said, "That's not where you should be looking". And he gradually moved into a much more semantic interpretation.
JM What did Bernstein have that you didn't have from Malinowski or Firth? They both have context, haven't they?
MAKH I think he added a coherent theory of social structure. I know he himself has now disclaimed some aspects of this but at the time, as it influenced me, he added a whole interpretative framework which enabled you to show not only the Whorfian effect, but also why patterns of educational failure were essentially class linked. In a society like the current western societies with their very strong hierarchical structures of class primarily and all the others, he asked "How were these, in fact, transmitted, maintained? What essentially is the nature of these hierarchies as semiotic constructs?" Bernstein put that in. I don't think that was there before. At the time there was all this stupid argument - Labov was trying to demolish him. But, if there was one person that needed Bernstein to give him theoretical underpinning, it was Labov. I mean, Labov doesn't make sense unless you've got something like Bernstein behind him.
GK We have a question about semiotics and systemics. Your model of language has connections to the work of Saussure and Hjelmslev alongside Firth. How would you position yourself in respect to continental structuralism and what role do you see for systemic theory in relation to post-structuralism and semiotics?
MAKH We need another seminar on this one. I mean, it's a good thing we didn't start with this question.
Firth, as you know, was very critical of Saussure on a number of points and regarded him as somebody who was perpetuating certain ideas in the history of Western thought which he didn't like, certain basic dualities. Now Ruqaiya would say, I think, that he was misrepresenting Saussure in a number of these ways, and maybe he was. In any case it seems to me that the world after Saussure was different from the world before. That's a fact and I certainly belong to the world after, although certainly there were things in Saussure, when I first read him as a student many years ago, and re-reading subsequently, that I wouldn't accept. I do think I share Firth's suspicion of langue/parole, although from a somewhat different standpoint.
As I see it, if you take the Saussurean view then you find it very difficult to show how systems evolve. But, it seemed to me that Hjelmslev had, to a certain extent, built on Saussure and also corrected that point of his; Hjelmslev's notions were much more adequate. To the extent that Hjelmslev differed from Firth, there are two important respects in which I would follow Hjelmslev. One is that Hjelmslev did have a very strong concept of a linguistic system, but a non-universalist one. This lies between the Firthian extreme which is: "There's no such thing as a language; there's only text and language events.", over to the other extreme of the universalists. Hjelmslev lies in the sort of middle position, which I think I would share. And then, of course, Hjelmslev constructed a fairly clear, useful, stratificational model. I haven't used it in the Hjelmslevian form and there are certain parallels built in between the different planes which I certainly wouldn't follow.
Certainly in the attempts to construct an overall pattern at the time when I was first doing this, I was very much influenced by Hjelmslev, and that's something which Firth just didn't have. In the last five to seven years I just haven't kept up with all semiotic and post-structuralist literature, so I've got a very partial picture. I was in Urbino for two or three summers in the early seventies, late sixties. That was when I first interacted with semiotics in the continental sense. It seemed to me that the general concept embodied in semiotics was a very valuable one because it enable me to say: "Here is a context within which to study language". Partly it's simply saying: "OK. We can look at language as one among a number of semiotic systems". That's valuable and important in itself. That then let's us look at its special features. We can then ask questions about its special status - the old questions about what extent language is unique because of the connotative semiotic notion - because it is the expression through which other semiotic systems are realised. And then thirdly at a deeper plane, semiotics provided a model for representing human phenomena in general, cultures and all social phenomena as information systems. This, of course, is really a development in line with technologies it seems to me. It goes with an age in which most people are now employed in exchanging information rather than goods and services.
Technology has become information technology. So our interpretations of the culture are interpretations as an assembly of information systems. This is what semiotics tried to interpret and increasingly, as I've mentioned, the physical sciences are interpreting the universe as an information system. So semiotics should provide a good home within which linguistics can flourish in this particular age, it seems to me. Now there are certain respects, of course, in which it's gone off in directions that I don't find so congenial.
JM If you have a well articulated comprehensive Halliday/Bernstein model, would that be an alternative to what the Europeans have in mind? With respect to the language and ideology conference last year and the way people were talking about ideology and language, it struck me as another way of talking about things that that Halliday-Bernstein model would be interested in. It's not doing something else. Gunther should really follow this up.
GK I feel that systemic theory provides the most worked out model for thinking about semiosis. And semiotics on the other hand has the ability to ask certain kinds of questions, or have a slightly different view point to look at language again. I think that's the formula of the relationship.
RH One of the problems of course is what is one thinking of as an example of post-structuralism.
RH If you're thinking of Derrida, that raises a different question which, at the deepest level, is really a question of realisation - the signifier and the signified and the relation between them. If you look at Bourdieu then that is a different question again and that question is the question of langue and parole, the sorts of relations that there are.
MAKH Bourdieu would be much more compatible with what Jim is referring to as the Halliday/Bernstein thing.
RH Yes. Greimas is yet another voice. He's not exactly what you would call a post-structuralist. But it is really very difficult with Barthes and Greimas to say exactly at what point they cease to be seen as structuralist. I myself find it very difficult to define the term structuralism. And that's what makes that question a little bit difficult to answer in one go.
MAKH We need another seminar on this one too. It seems to me, that in so far as post-structuralism has become a literary theory, then some of the ideas that are used in discussions of literature and are ascribed to structuralists by people working in the general semiotic and post-structuralist field really aren't there at all. I mean they're quite different from what these people are actually saying.
RH That's generated a very interesting point: how it is that a discipline retains it's old assumptions while using new names, and resists any innovations. Literary criticism is one of the disciplines that is a prime example of this kind of thing. One should study that for how to retain ideology and not to change it.
GK It seems to me, just to make two comments, that structuralism and post-structuralism asks questions of linguistic theory which are important to ask. Derrida's work, for instance, really sharpens up the question about system, because it in itself is a model that works without system. It works only with the surface effects of structures. So it asks really important questions about system. But the thing that interests me most is that post-structuralism asks questions about the constitution of language uses, in linguistic terms, which linguistics, because of its concerns with the system itself, hasn't I think addressed as fully as it might. That seems to me important.
Anyway, we have a question on speech and writing. Is there an implied valuation of speech over writing in your descriptive work. The second part of the question which is: How does your recent work on grammatical metaphor relate to this issue?
MAKH In a sense there is an implied valuation of speech over writing in relation to this notion of levels of consciousness, if you like. It seems to me there's a very important sense in which our whole ability to mean is constructed and developed through speech, and that this is inevitably so. In other words speech is where the semantic frontiers are enlarged and the potential is basically increased. I know that one of the problems here is that there's a risk in this being interpreted like the old, early twentieth century structural linguists, who insisted on the primacy of speech over writing for other reasons. But there are things I want to say about natural spontaneous speech which do, in a sense, give it a priority.
This has been partly of course political because I feel that it is essential to give speech a place alongside writing in human learning and therefore in the educational process. I still feel very strongly about that. Now the work on grammatical metaphor is partly an attempt to explore the nature of the complementarity between the speech and writing. There are modes of action and modes of learning which are more spoken, speech-like and which are more naturally associated with spoken language, and others which are more naturally associated with written language. This is something which needs to be explored. I'm always asking teachers if they feel that there are certain things in what they do which are more naturally approached through the spoken. At a deeper level differences between speech and writing have to be explored in the wider semiotic context that we're working with.
We need to ask about writing as a medium, the development of the written language, and the development of technical discourse, exploring a technicalisation that is part and parcel of the process of writing and which involves grammatical metaphor. We need to ask what the nature of the realationship among these things is and between all of these and the underlying sorts of phenomena that they've used to describe.
Beyond this it's the whole question of how far can we use notions of grammatical metaphor, and indeed the whole systemic approach to language, to try and understand the nature of knowledge itself. It relates to what we've been talking about in some of these seminars on a language-based theory of learning.
When I started in the E.S.L. area I remember going to Beth Ingram, the psychologist at the
We had a lot of useful ideas but nothing that could be thought of as a general learning theory into which this our work could be fitted. So it seems to me we have to ask the question "Well, can we build one out of language" I mean "Don't we by now know a lot more?" I am obviously influenced by Jim here who's been pointing out all along that linguistics should in fact simply take over a lot of these things and see what it can say from a language point of view. And I certainly think that we have to work towards a much more language-based theory of learning and language-based theory of knowledge. And in that notions like grammatical metaphor, and the difference between spoken and written language, are obviously fundamental.
GK Our next question in a sense addresses that in a somewhat broader way. Your work has paved the way for a radically larger role for linguistics in the humanities and social sciences and perhaps beyond than has been possible in the past. What, to your mind, are the limits of semiosis? Just how far can a language-based model be pursued before turning over to other disciplines?
MAKH I think that we've drawn disciplinary boundaries on the whole far too much. We had to have them of course. I think Mary Douglas sorted that one out many years ago very very well. The discourse, so to speak, had to be created in definable circumscribed realms. But the cost of this was defining these far too much in terms of the object that was being studied. Thus linguistics is the study of language, and so on. Now that is really not what disciplines are about. A discipline is really defined by the questions you are asking. And in order to answer those questions you may be studying thousands of different things. Linguists start by asking questions about language. And if you ask "Well how far do questions about language take us?" , then the answer is "They take us way beyond anywhere that we are yet operating in." The frontiers are well beyond. I don't know where they are, but they're certainly well beyond where we are at the moment. They certainly take us into a lot of questions that have been traditionally questions of philosophy, which has always been about language to the extent it's been about anything and into questions of general science. I mean, this is why I've become increasingly interested in scientific language and general problems of science.
It has become increasingly clear that you can ask questions about language which turn out to take you into and even way beyond human systems. So I don't know where the frontiers are but they're certainly a great deal further than I think we've been able to push them. And, in a sense, I've tried to have this kind of perspective in view all along; I wanted a linguistics which is not defined by object language as object rather by questions. These questions began by being questions about language but eventually expand into areas that we don't expect. I certainly think that we should be fighting a lot more for the centrality of linguistics, not only in the human sciences but in science generally, at least for the foreseeable future.
GK In what way do you mean that? As a means of elucidating what scientific disciplines are doing?
MAKH Yes. Current thinking has been emphasising the similarities among human, and between human and non-human systems, between human and physical systems if you like. Take first of all Lemke's work, which I think is tremendously important, on dynamic open systems. He's taken over the social semiotic notion, which he's characterised these essentially physical terms. Language fits in, but then becomes a way of looking at other human semiotic systems, which are language-like in this respect and for which language serves as both the semiotic which realises them (the connotative semiotic sense) and also a model and a metaphor in a very important sense. I think you can go beyond that now into physical systems. The universe in modern physics in being thought of as one, whole, indivisible and conscious. In other words the present generation of physicists is adding consciousness to the universe, talking about exchange of information.
That came originally out of quantum physics. Now my point is I want to say not "one, whole, indivisible and conscious" but rather "one, whole, indivisible and communicative". In other words I want to say the universe, in an important sense, is made of language, or at least made of something of which human language is a special case. Taking the notion of a natural grammar, one step further is to say that language is as it is because it not only models human semiotic systems (realities we construct in a very important sense); it also models natural systems. Obviously, talking like that is talking in a very abstract way; but on the other hand, I think that there is an important sense in which the situation has been reversed. Instead of modelling all our thinking in some respects on physics, as in the classical period (and from physics via biology it got into linguistics), I think there's an important sense in which in the next period the thing is going the other way round. We are going to start from the notion that the universe as a kind of language if you like, and therefore move outwards from linguistics. Towards human and then biological and then physical systems.
GK A materialist linguistics.
GK We have one last question which is about linguistics and machines. Very early in your career you worked on machine translation and since then your work has played a central part in a number of artificial intelligence projects. Is this because of or in spite of your socio-functional orientation? How has your recent involvement in I.S.I. influenced your thinking about language, linguistics and machines?
MAKH I don't see the interest as in any sense conflicting. As you know I have never thought of either the machine or the linguistic theory as in any sense a model of human psychological processes, so there's no question of seeing some model of the brain as a common base.
Now I've had one concern throughout which is that it seemed to me right from the beginning when I first tried to learn about this back in the late fifties in the Cambridge Language Research Unit that the machine was, in principle, a valuable research tool. Now that was the nature of my first interest. By seeing if we could translate Italian sentences into Chinese, which we were doing at that time, we learned more about language. I've been in and out three times now. First of all, while in the very early stages, we had some fascinating discussions and it was all great fun; but it was obvious that the technology itself was still so primitive that we were constrained by the hardware the internal housekeeping rules so that we weren't actually learning anything about language in the process.
I had another interest in it which is that I felt that machine translation had an important political role to play. There were lots of cultures around the world where people were beginning to be educated in a mother tongue and if you could possibly have a machine to translate a lot of text books at least it would help the process. So there are practical concerns like that. Then in the late sixties I came back again with the project on the Linguistic Properties of Scientific English that Rodney Huddleston and Dick Hudson, Eugene Winter and Alec Henrici were working on. Henrici was the programmer and at that time we used the machine to do one or two things in systemic theory. For example he had a program for generating paradigms from a network. So you could test out a network that way. And he could even run little realisations through it. But again there were tremendous limits in the technology. At that time I started being interested in generalising and parsing programs.
I wanted to test the theory and of course, I was responding to external pressure. At that time in the sixties unless you could show that your theory was totally formalisable it was of no interest and I was responding to these pressures. This was why I was interested in Henrici actually generalising clauses by the computer.
But my real interest in that was that I was beginning to realise that you could no longer test grammars except with a machine, in the sense that they were too big. If you really had delicate networks, the paradigms were just huge; you had to have some way of testing this.
There was still a limit on the technology then. I wanted to write the grammar in metafunctional terms. I wanted to say "I don't like the sort of transformational model where you have a deep structure and then obligatory transformations and then optional transformations on top of them. I want to be able to represent things as being simultaneous and unordered". And the answer was "Well, we can't compute this and therefore it must be wrong". I never accepted that answer. It always seemed to me to be incredibly arrogant to say that if our logic or our hardware can not do something at this stage therefore it must be wrong. So I just backed off again and I never thought I would come back into it at all.
I thought that was it until Bill Mann came along when we were in
Parallel processing is not a problem but there are still constraints on the extent to which each of these processes can consult all the others as it's going along and modify it's own behaviour in the light of that consultation. It seems to me that if you can get that kind of thing available then we really can learn a lot by constructing parsing and generating programs and using them to test the grammar. It's been as a research tool mainly that I have been interested in this although there obviously are practical applications that are useful.
RH Where is the point where systemics needs more growth? Which direction is it going? What is your hope that systemicists would develop?
MAKH Well, more of what they are doing I think. I mean we just need more people, more time, more resources, the usual thing. One of the things that we have been very weak on is any kind of clinical applications and the underlying theory that goes with those. Bruce MacKellar is the one who has certainly done most that I find interesting, but he hasn't written up much of it yet on that side. I mean, he's written an enormous amount of background material, but less about the neurolinguistics. McKellar's notion is that systemic theory is likely to be useful, moreso than others he thinks, in developing neurolinguistics. He doesn't believe that there is such a thing yet, but he sees ways of doing it. And the interesting thing is that he sees this not so much in relation to the particular representations of the grammar on the linguistic system, as in the social semiotic background to it. Now that's one development I think is very important - towards a neurolinguistics and towards clinical applications. Again we will in turn learn from these things. So I would like to see it far more used in context of aphasiology and all kinds of studies of developmental disorders.
RH Let's go back to the machines and how they can be used for testing the grammar. At the present moment all they can do is test the grammar of a clause or with luck, clause complex; but they are not able to do anything yet on what constitutes a normal natural sequence of people's sayings in any context of situation without going up and building in context of situation. That was the context in which I had raised that question of probability because it seems to me that the only way that probability is going to link up with text is in some through context.
MAKH Well, there has to be some sort of register model, as part of it. But I don't know that in principle there's any reason why this can't be built in, given that point that I was saying about the remaining limitations on the technology. The environment, as they call it in the I.S.I. project, which means the knowledge base and the text planner, are still very primitive. But they're primitive because we just haven't had enough people doing enough work on them. I think that given a research effort in that area then it should be possible to represent these things in such a way that they can be part of the text generation.
 The questions in the interview schedule were designed by Hasan, Kress and Martin and given to Halliday a few days prior to the interview. Hasan and Martin subsequently edited the interview into its present form.
 Horvath, a Labovian sociolinguist, was Halliday's first appointment to the Department of Linguistics he founded at the
 [MAKH to unpack...Further Education Training Scheme???]
 Firth, J R 1950 Personality and language in society. Sociological Review 42. 37-52. [Reprinted in J R Firth 1957. Papers in Linguistics 1934-1951.
 Halliday did in fact join the Communist Party, and was a member until 1957 when he left over the party's failure to condemn or even to properly discuss condemning
 Halliday's Ph.D. thesis was published as Halliday, M A K 1959 The Language of the Chinese "Secret History of the Monguls".
 For an overview of Firth's theory see Firth, J R 1957b A Synopsis of Linguistic Theory, 1930-1955. Studies in Linguistic Analysis (Special volume of the Philological Society).
 Allen, W S 1956 System and structure in the Abaza verbal complex. Transactions of the Philological Society. 127-176.
 Halliday, M A K 1964 Syntax and the consumer. C I J M Stuart [Ed.] Report of the Fifteenth Annual (First International) Round Table Meeting on Linguistics and Language Teaching.
 Halliday's first scale and category grammar of English was written on the cardboard inserts he received inside his shirts from
 Lemke, J. 1984 Semiotics and Education.
 For a summary of the discussions on synoptic and dynamic representations referred to here see Martin, J R 1985 Process and text: two aspects of semiosis. J D Benson & W S Greaves [Eds.] Systemic Perspectives on Discourse vol. 1: selected theoretical papers from the 9th Internatinal Systemic Workshop.
 See Halliday, M A K 1975 Learning How to Mean: explorations in the development of language.
 For further discussion see Halliday, M A K 1987a Language and the order of nature.
 [MAKH to fill in reference...]
 For a summary article see Halliday, M A K in press Towards probabilistic interpretations.
 For a retrospective overview of this initiative which produced the Breakthrough to Literacy and Language in Use materials see Pearce, J, G Thornton & D Mackay The Programme in Linguistics and Engish Teaching,
 See Bernstein, B. 1971, 1973, 1975, 1990 Class, Codes and Control, Vols. 1, 2, 3, 4.
 Halliday, M A K, A McIntosh & P Strevens 1964 The Linguistic Sciences and Language Teaching.
 During the late 1970s the Curriculum Development Centre in
 This is a fluid network of linguists and educators (anchored by Fran Christie and initiated by Halliday in 1979) which has held several conferences on language in education issues around
 See Gray, B 1985 Helping children to become language learners in the classroom. M Christie [Ed.] Aboriginal Perspectives on Experience and Learning: the role of language in Aboriginal education.
 See Halliday, M A K 1967 Intonation and Grammar in British English.
 See Halliday, M A K in press New ways of meaning: a challenge to applied linguistics. to appear in
 For discussion of these debates see Atkinson, P 1985 Language Structure and Reproduction: an introduction to the sociology of Basil Bernstein.
 See Hasan, R 1985 Meaning, context and text: fifty years after Malinowski. J D Benson & W
 For a recent statement on levels see Halliday, M A K in press How do you mean? M Davies & L Ravelli [Eds.] Papers from the Seventeenth International Systemic Congress, University of Stirling, July 1990.
 For the proceedings of this conference see Threadgold, T, E A Grosz, G Kress & M A K Halliday 1986 Language, Semiotics, Ideology.
 See Halliday, M A K 1985a An Introduction to Functional Grammar.
 See Halliday, M A K 1988 On the language of physical science. M Ghadessy [Ed.] Registers of Written English: situational factors and linguistic features.
 See for example Martin J R 1986a Grammaticalising ecology: the politics of baby seals and kangaroos. T Threadgold, E A Grosz, G Kress & M A K Halliday. Language, Semiotics, Ideology.
 [MAKH to add reference...???]
 The Information Sciences Institute in
 See Halliday, M A K 1962, Linguistics and machine translation. Zeitschrift fur Phonetik, Sprachwissenschaft und Kommunikationnsforschung. 15.1/2. 145-158. [republished in M A K Halliday & A McIntosh. 1966. Patterns of Langugae: papers in general, descriptive and applied linguistics.
 See Huddleston, R D, R A Hudson, E Winter & A Henrici 1968 Sentence and Clause in Scientific English.
 See McKellar, G B 1987 The place of socio-semantics in contemporary thought. R Steele & T Threadgold [Eds.] Language Topics: essays in honour of Michael Halliday.