LHS (print) issn 1742-2906

Linguistics and LHS (online) issn 1743-1662

the Human
Sciences
Article

On matter and meaning:

The two realms of human experience

M. A. K. Halliday

Abstract
There are two phenomenal realms that we as human beings inhabit: a world of matter,
and a world of meaning. Both matter and meaning are involved in all the regions of
our experience. Meaning relies on matter to make it accessible to a receiver; and matter
relies on meaning to organize it. Processes that take place in human consciousness
may be conceptualized as processes of meaning. Language is the leading edge of mean-
ing, even if not all types of human semiotic are necessarily realized by language. A
language is more than just a semiotic system, a system of meanings; it is also a system
that makes meanings, a semogenic system; and the source of this semogenic power is
grammar. The powerhouse of a language is its lexicogrammar, the unified stratum of
syntax and vocabulary; so thinking about meaning means thinking grammatically.
Expressed in functional terms, the grammar both construes and enacts: it enacts
the social process, our relationships one with another; and it construes the human
experience. Language can be seen as the prototype of a semiotic system; hence a theory
which is designed to represent the multidimensional ‘architecture’ of language should
be ‘thick’ enough, and rich enough, to offer insight into other semiotic systems. As
well as providing a metaphor for language, linguistics also stands as metaphor for
the whole of meaning as theorizing – for the ability of the semiotic realm to construe
itself into successive planes of virtual reality, in the (so far) unremitting human effort
to understand.
Keywords: history, language, meaning, semiotic system, lexicogrammar,
grammatics, metafunction

Contact
M. A. K. Halliday, Unit 14, 133 Sydney Road, Fairlight NSW 2094, Australia.

LHS vol 1.1 2005: 59–82

©2005, equinox publishing
60 linguistics and the human sciences

I
This paper is about the nature of happenings, and more particularly the hap-
penings that constitute the human experience. What happens becomes history;
someone who studies and tells what happened is a historian. I am not a historian
– I can enjoy reading history free of the sense of entanglement that goes with
reading in your professional domain; as a youth, my greatest feeling of discovery
came from reading Gordon Childe’s What Happened in History, and thanks
to my near-contemporaries such as Christopher Hill and Eric Hobsbawm I
have always been able to refresh that feeling when I needed to. ‘Telling’ means
selecting, ordering, interpreting, explaining, in some proportion or other; but,
as Hobsbawm puts it, ‘what historians investigate is real’ (1997: viii)1. Of course,
we can construct imaginary, ‘virtual’ happenings, such as those in dreams; but
even those are modelled on experience of the real.
As a student, many years ago, my encounter with history was with the history
of China; and since I am writing this in Hong Kong, where Chinese and English
history met, let me draw an example of happening, and telling about happening,
from these two sources. The motif is a familiar one: it is that of cycles, cyclical
patterns in the ordering of historical events. Chinese history has traditionally
been interpreted in terms of cycles, as each new dynasty first flourished, then
matured, and then declined, only to be overthrown by a successor which then
went through the same sequence of events. A typical cycle lasted about 250
– 300 years.
England likewise went through a number of cycles, as it tended to drift away
from the European continent and then get forcibly reabsorbed, in this case
about once every 500 years. First (in history) came the Celts, having travelled a
long way from their earlier home in the east; 500 years later the Romans arrived;
after another five centuries the Anglo-Saxons; and then, just one thousand years
ago, the Norman French. Each of these invasions reaffirmed England’s place as
part of Europe (as it had been physically before the ice melted); and in between
each it gradually moved away again, taking on a distinct identity of its own.
In the last 1,000 years this cycle has continued; but the invasions have taken a
different form: not physical force, but intellectual and political pressure. We can
see this happening today, with England being dragged, more or less reluctantly
(depending on one’s viewpoint), into the European Union, the Eurozone and
all the rest. The Eurocycle has come round once again. But there was one cycle
in between, 500 years ago, that Europeans called the Renaissance, the ‘re-birth’
of the spirit of inquiry when new forms of knowledge began to emerge; this
started in Italy but soon spread throughout Europe, leading on in England to
what came to be known as the ‘industrial revolution’.
 M. A. K. Halliday 61

It is well known that in China, already in the Southern Song dynasty, many
of the conditions prevailed that in Europe, several centuries later, formed the
point of departure for industrialisation: in textiles, there was large-scale factory
production, with advanced machinery driven by a reliable source of energy,
namely water power. Historians, both Chinese and foreign, have often asked
why China, which had been ahead of Europe in technology since Han times, did
not take this next critical step into the machine age; and various considerations
have been put forward to explain it (Needham, 1978–1995; Elvin, 1973). Or we
could ask the question the other way round: why did such a development take
place in western Europe? What were the circumstances that led to the evolution
of a capitalist economy and scientific – industrial technology?
Now I am obviously not going to suggest an answer. What concerns me here
– the reason for bringing up one of these questions of history – is the kinds of
explanation that may be offered. In the broadest terms, some of the discussion
has foregrounded the technological and other material conditions that obtained
at these places at these times; some has emphasised rather the belief systems,
social relationships, forms of discourse and the like. Needham many years ago
suggested a connection between the European idea of God as the ultimate agent
and the search for scientific explanations of material effects once God was no
longer being appealed to in this role. ‘Marxist’ explanations are often assumed
to be purely in material terms, based on some form of economic determinism,
although as Hobsbawm points out (1997: Chapter 13) Marx fully recognised
the importance of ideas as a historical force; he just insisted that they were
not independent of a society’s economic base. Most historians would seek an
explanation in some combination of the two: the material and the – let us say
– non-material historical conditions.
But what are the ‘non-material’ conditions? They can be thought of in ideo-
logical, religious, social, psychological, cognitive, cultural and no doubt other
terms as well; there are many headings under which they can be described
and usefully explored. But to what extent does the non-material constitute
a coherent, unified domain of human experience? I think there is a unifying
concept, an underlying factor that is present in all such manifestations: namely,
meaning. The two phenomenal realms that we inhabit, as human beings, are the
realm of matter and the realm of meaning. Human history is the unfolding of
a constant interplay, and a constant tension, between these two.
From matter, we can access the adjective ‘material’. From meaning we cannot
derive any adjective; so we have to go to Greek and call it ‘semiotic’. The two
realms that we inhabit, then, are the material and the semiotic; and both are
involved in all the regions of our experience.
62 linguistics and the human sciences

II
I have said I am not a historian; nor, I must add, am I a philosopher. By training,
I am a grammarian – by training, but also by inclination: I tend to think gram-
matically, especially where I meet problems that need to be solved. This does
not mean that I am going to suggest a grammatical explanation for the above
historical problem. I do not believe that, if the scientific – industrial revolution
took place in a part of Europe rather than in a part of China, despite similar
technological conditions, this had something to do with the differences between
Chinese grammar and the grammar of English or other European languages.
Every language has the potential of being the vehicle of scientific thinking and
theorising; and Chinese had already gone a long way along this road – in 1200
A.D. probably only Arabic, Latin and Sanskrit had travelled so far.
Of course, every language realises this potential in its own way, and certain
steps may require less expenditure of grammatical energy in one language than
in another, and at one stage in a language’s history than another – although
this is highly likely to balance out in the long run. For example: if we compare
modern Chinese with modern English, as used in the discourses of science,
then (1) technical terms, and technical taxonomies, require less grammatical
energy in Chinese than in English; (2) compound nominal expressions require
less energy in English, while (3) logical (tactic) sequences take up about the
same amount of energy in both (Halliday, 1993). We can reject any suggestion
that there is something about the grammar of English (or Dutch, or German,
or French) which enabled speakers of these languages to construct scientific
theories and steam engines; or something about the grammar of Chinese
which inhibited Chinese scholars and engineers from advancing along the
same paths2.
In what sense, then, does grammar come into the picture? Not, of course,
the image of ‘grammar’ that many people carry around with them (including
some who are sophisticated, theoretically-minded scholars in other disciplines).
This is still often the grammar of the primary school, where they were taught
that a language is a set of rules to be obeyed, whether in speaking and writing
their mother tongue or in struggling to master a second or a foreign language.
This is rather as if my image of mathematics was struggling to prove that the
square on the hypotenuse of a right-angled triangle was equal to the sum of the
squares on the other two sides, with no idea of why I was being required to do
so. A language is a highly complex system – perhaps one of the most complex
in the known universe. This is what makes grammar such a powerful tool for
thinking with. But by the same token, in investigating how language works we
are inevitably drawn into abstract realms of thinking and reasoning, requiring
a technical discourse with its associated conceptual taxonomies. People often
 M. A. K. Halliday 63

resist technical terminology about language, even though their primary school
mathematics and science was already loaded with batteries of specialised terms
– which cause no serious problem to the brain of an eight-year-old child.
A language is a system of meanings: a semiotic system, to give it a technical
name. But it is more than that; it is a system that makes meanings: it is not only
semiotic but semogenic. There are many systems of meaning in our lives, but not
all of them are meaning-creating. A system of railway signals, for example, or
traffic lights, is semiotic, but not semogenic. There are other semiotic systems
which do create meaning: forms of visual art and music, for example; but prob-
ably language is unique in the extent of its semogenic power. I emphasise this
because, again, like the notion of grammar, it is not in harmony with received
notions about language. The usual way we talk about language is by saying that
language ‘expresses’ meaning, as if the meanings were already there – already
existing, in some formation of other, and waiting for language to transpose
them into sound, or into some kind of visible symbols. But meaning is brought
about by language; and the energy by which this is achieved, the source of its
semogenic power, is grammar.
One of the sources of complexity is that there is more than one kind of mean-
ing in a language; so the grammar is doing more than one job at once. Expressed
in functional terms, the grammar both construes and enacts: it enacts the social
process, our relationships one with another; and it construes the human experi-
ence. We refer to these in systemic functional linguistics as metafunctions: the
interpersonal metafunction and the experiential metafunction3.
Here I want to consider meaning in its experiential aspect: meaning as the
construction of human experience. Human experience is extraordinarily
rich and varied, mediated through different senses on various levels; what
the grammar does is to transform this experience into meaning. Since all
human beings live on the same planet, and all have the same brains – the same
neurophysiological make-up – there is much in common to the ways in which
this transformation of experience into meaning is brought about, in every
language. But there is also room for variation; some of it motivated by the dif-
ferent ecosocial environments of different human groups, some of it arising out
of the self-organising strategies of the system, and some of it simply random.
Our interaction with our environment is so complex and multidimensional
that there has to be a lot of ‘play’, or indeterminacy, in the construal for it to be
able to work at all.
What does it mean to say that the grammar ‘transforms experience into
meaning’? Another way of putting this would be to say that grammar is a theory
of human experience. The lexicogrammatical representations, or wordings
(I should make clear that ‘grammar’ includes vocabulary, the more specific
elements of the wording) – that is, the bits of discourse that we recognise as
64 linguistics and the human sciences

speech, or as writing, construe fragments of experience, real or imaginary, in

ways which relate each instance to this overall theory. So if someone says to
me, as they did a few weeks ago, That window was blown open by the typhoon
(or, since it was actually in Chinese, Nage chuangzi gei dafeng chuikaile), this
is not simply a record of an event, observed or narrated – not simply an act
of annotation; it is an act of interpretation, a construal in terms of the theory,
which is a complex construction involving processes and agency and shared
(or, again, imagined) visual space, of phases and aspects in time, of entities
and forces and the categories to which these are to be assigned, and so on. It
thus relates this particular phenomenon to innumerable other phenomena in
innumerably many ways.
So the reason that one can say with confidence, as I did earlier on, that every
language has the full potential for scientific reasoning and theorising, was
that every language in its ordinary everyday grammar already incorporates
a highly complex theory of human experience. The meanings construed in
wordings like blown open and was and by, or the Chinese chuikai and le (liao)
and gei, or even window and typhoon, are part of a very sophisticated model of
ourselves and our environment; such a model did not emerge all at once, but
evolved over the ten thousand or so generations of homo sapiens – who in turn
inherited the foundations for such a model from an even remoter ancestry in
the hominid line.
Once given this construal we can then retheorise it at a different level, referring
to materials and structures and forces, all of which can be accurately measured,
and computing the mathematical relations among them. This is what we call a
‘scientific’ theory, as distinct from the commonsense theory that is enshrined
in our everyday grammar. A scientific theory is also the product of long evolu-
tion, although in this case heavily subsidised by conscious design; whereas the
commonsense theory operates below the level of conscious awareness – that
is, until you become a grammarian. The elements of a scientific theory are of a
more abstract nature – they may be virtual entities, existing only on the semiotic
plane and with no direct analogue in the material world; but such a theory is
still ultimately the output of the semogenic power of the grammar.

III
In formulating the title of this paper, I had in mind a quotation from a discus-
sion by the biologist George Williams, where he said (1995: 43):
Evolutionary biologists… work with two more or less incommensurable
domains: that of information and that of matter…. These two domains
will never be brought together in any kind of the sense usually implied by
 M. A. K. Halliday 65

the term ‘reductionism’. You can speak of galaxies and particles of dust in
the same terms, because they both have mass and charge and length and
width. You can’t do that with information and matter. Information doesn’t
have mass or charge or length in millimetres. Likewise, matter doesn’t have
bytes. You can’t measure so much gold in so many bytes. It doesn’t have
redundancy, or fidelity, or any of the other descriptors we apply to informa-
tion. This dearth of shared descriptors makes matter and information two
separate domains of existence, which have to be discussed separately, in their
own terms.
I have cited and commented on this passage elsewhere (Halliday, [1998] 2004).
The key point here is the notion of two distinct phenomenal realms, each an
essential component of the human condition and neither reducible to the other.
They are defined, for Williams, by the properties whereby they can be measured:
matter by mass, density, extension in space; information by bytes – kilobytes,
megabytes, gigabytes and whatever may come after.
I have a problem with this notion of information: not with the phenomenon
itself rather, or with its measurability in bytes, but with the notion that informa-
tion defined in this way is an ultimate phenomenal realm on a par with the
realm of matter. I want to replace ‘information’ here by ‘meaning’: the two
realms of matter and meaning – the material and the semiotic. This is not just
because I think like a grammarian, though that is obviously a factor in my
underlying rationale; it is because I don’t think all meaning can be measured.
There is no doubt that some kinds of meaning, or at least some manifestations
of meaning, are measurable; it seems sensible to reserve the term ‘information’
for just that part of the realm of meaning that can be measured in bytes – or
in terms of information and redundancy (which incidentally can be used very
profitably in quantifying grammatical meaning (Halliday & James, 1993). But
it is not, I think, an essential property of this second realm that it can always be
expressed in quantitative terms. This is why I prefer the more general concept
of ‘meaning’: matter and meaning, rather than matter and information.
These are the phenomenal realms which we as human beings inhabit. We
inhabit a world of matter, and we inhabit a world of meaning. To return to the
historical perspective with which I started: at any given moment in history,
the environment of any human group – the ecosocial context in which they
are located – is some particular intersection of these two realms, the material
and the semiotic. This environment is at once both enabling and constraining:
it contains within itself both a potential for action and a limitation on what
actions are possible. Perhaps if we accumulated enough data we could express
these in terms of probabilities. Neither the material conditions alone nor the
semiotic conditions alone are determining; it is the interplay between these
66 linguistics and the human sciences

two which defines the human situation – whether for the individual, the social
unit, the state, or for the human race as a whole.
China had developed movable block printing five centuries or so ahead of
Europe. By the time of the Yuan dynasty Chinese scholars had developed all
the ‘styles of scientific thinking’ (identified by Crombie, 1994) that evolved
in Europe, with the exception of the probabilistic mode (Elvin, 2002). But, as
Elvin pointed out, there was a weakness in their recording and transmission
of knowledge: Chinese scholarship lacked the dimension of continuity, of suc-
cessive development, such that meanings could be seen to accumulate through
time. We may contrast this with the strong sense of succession that came about
in Europe, with the chain of knowledge linking Copernicus, Kepler, Tycho
Brahe, Galileo and Newton. But we might also note that for a thousand years
before this Europeans had largely failed to build on the learning of classical
Greece; it was the Arab scholars who picked this up and ran with it. Clearly
the current systems of thought and belief, and the forms of social interaction,
all play a part; these too are ways of meaning – ways of exchanging meaning,
since semiotic processes, like material processes, are essentially processes of
exchange. And the semiotic systems that underlie them, like material systems,
are both enabling and constraining, and show the same tendency to shift from
one to the other over time. Ideologies which start by opening things out end
up by closing things down.
But I have been taking the notion of meaning, of semiotic systems-and-proc-
esses, rather much for granted; and it needs to be problematised and discussed.
Probably people of all cultures are aware of a non-material plane of existence,
which has its own parameters and its own hierarchies of value. This awareness
takes a wide variety of institutional forms, from esoteric mysticism at one end
of the scale to a code of social morality at the other; and it is construed in terms
of various dualities such as body and soul, or matter and mind, physical and
mental; and at a more theoretical level in schemata such as materialism versus
idealism. We can investigate this sensibility grammatically, in terms of the
grammar’s theory of process types; when we examine these we find that many
languages make a clear grammatical distinction between material and non-
material processes, the non-material being construed as processes of (typically
human) consciousness – knowing, thinking, feeling – in which material entities
are not accepted as the site where such a process takes place. There are often
other differences as well: processes of consciousness have different relations
to time, and different event structures, from processes of the material kind.
Like almost all significant features of our everyday grammars, these distinc-
tions lie well below the level of people’s conscious awareness. Matthiessen has
argued convincingly that our familiar concepts like ‘mind’, and their learned
counterparts like ‘cognition’, derive ultimately from this duality that is built into
 M. A. K. Halliday 67

the grammar itself, and that is reinforced for us every time we talk or listen to
language (Matthiessen, 1993; 1998). In other words, we are inclined to think in
such dualities precisely because that is how our languages are organised. It made
sense, in the grammar’s construction of experience, for material and semiotic
events to be construed as two significantly different kinds of process.

IV
So it should come as no great surprise that we think in terms of such dualities:
our grammar construes the duality for us. Such patterns are bound to remain
subconscious; we could not talk if we were aware of them all the time. But if
we listen to small children learning language – learning their first language, the
mother tongue – we can observe how they get constructed, and then buried
from sight. Such a dichotomy began by being enabling: it helped us to interpret
and so survive in our ecosocial environment. But it is also not surprising if it
has now become a constraint. I have written elsewhere ([1990] 2003) about how
this kind of dualistic picture, of a sharp division between conscious and non-
conscious beings, is now conspiring to make us destroy that environment. And
at the level of designed theory, many scientists and philosophers are finding this
dualistic habit of thinking now quite seriously dysfunctional (Rose, 2003).
But let me rather try to recontextualise it, in the terms of matter and meaning.
I don’t think we can regard processes that take place in human conscious-
ness as ‘information’; but I think we can conceptualise them all as processes
of meaning. Not all processes of meaning end up as wording; that is how
language organises them, as lexicogrammar (words-and-structures); but some
are organised in other ways, and others perhaps not organised at all4. Some
processes that we interpret as meaning are not, in fact, human: we recognise
that other creatures also know how to mean, in their own terms; some direct
their meanings at humans, like dogs and horses, and some, like birds and many
orders of mammals, communicate with each other. But there are numerous
types of human semiotic that are also not made of language: music, dance,
the visual arts; ritual, and semi-ritualised forms of behaviour, like clothing;
and images and graphic representations of all kinds – tables, charts, graphs,
diagrams, maps, logos and so on.
These human semiotics all depend on language to some degree or other,
ranging from literature, which is made of language though having its own
modes of meaning and values, to music, perhaps the least dependent of all
but still ultimately reliant on the factor that those who practise it also traffic
in language. But what links all these processes together, with each other and
also with language, is that they all depend on physical processes in order to
come into being. These may be sound waves travelling through some medium,
68 linguistics and the human sciences

usually air; or light waves together with the substances from which light is
being reflected – building materials, paint, paper, computer screen and so on.
Meaning relies on matter to make it accessible to a receiver; in linguistic terms,
meaning depends on matter to realise it. A semiotic system is made of meaning;
but to be realised, as process, it has to materialise – to become matter.
In that respect, however, a semiotic system is no different from systems of
other kinds: biological systems, and social systems. These are also constructed
of matter; they are subject to the laws of physics, like everything else. But they
raise an interesting question of ordering. A social system is also necessarily
biological – but not the other way round. Let us start with a physical system.
A physical system is just that: a physical system. What is systemised is matter
itself, and the processes in which the system is realised are also material. But
a biological system is more complex: it is both biological and physical – it is
matter with the added component of life; and a social system is more complex
still: it is physical, and biological, with the added component of social order,
or value. So then a semiotic system is still one step further in complexity:
it is physical, and biological, and social – and also semiotic: what is being
systemised is meaning. In evolutionary terms, it is a system of the fourth order
of complexity.
Now it may be objected: is this really a further step? Surely there can be
no social order without meaning? True. But in that sense, can there exist any
biological order without meaning? or even physical order? I have said that
meaning needs matter to realise it; perhaps we should cap this by saying that
matter needs meaning to organise it. This would be saying that all organisation,
all departure from a purely random state, is a form of meaning. A physical
system, then, is not just a physical system – or rather, it is not just made of
matter; being a system, it is also made of meaning. This is a reasonable view,
although it appears to extend the meaning of meaning beyond the notion of a
semiotic system as I have been describing it.
The trick lies in the word system, or in the slippage between system and phe-
nomenal realm. If we characterise a system, say a social system, we are saying
where the organisation is taking place, or rather, perhaps, what it is that is being
organised; and in that sense we can justifiably say that a language is a system
of a fourth order of complexity. First, it is transmitted physically, by sound
waves travelling through air; secondly, it is produced and received biologically,
by the human brain and its associated organs of speech and hearing; thirdly,
it is exchanged socially, in contexts set up and defined by the social structure;
and fourthly, it is organised semiotically, as a system of meanings – a meaning
potential, as I have called it. We cannot leave out any of these components, if we
want an explanatory model of language; they are all involved in the semogenic
enterprise. Not in an ideal harmony, of course; there are conflicts, disjunctions,
 M. A. K. Halliday 69

compromises within and between them, as there are bound to be in any such
complex organisation; but every component is a necessary ingredient of the
whole.
Now it may be that all these different kinds of complexity, different forms
of systemic order, will eventually be explained by the same methodology that
gave birth to modern physics. Physical scientists can reasonably claim that the
power of such reductionist strategy has already been demonstrated in the case
of biological systems, via biochemistry, molecular biology and genetics (Wilson,
1998). But there is still a very long way to go. If we set up a continuum between
the two phenomenal realms, matter and meaning, with the intermediate points
defined according to the mix, then physical systems lie towards the ‘matter’ end:
meaning comes in only when we define all systemic organisation as meaning.
Semiotic systems lie towards the other end, the ‘meaning’ end: matter comes
in just in the processes whereby the meanings are realised. But there can be no
‘pure’ categories, at either end, because the two realms must always interpen-
etrate: as I put it just now, meaning needs matter to realise it, and matter needs
meaning to organise it. In this perspective, language ultimately evolves out of
the physical organisation of matter: it is an indirect outcome of the big bang.
Physical systems provide the prototype for exploring the complexity of our
material environment, while semiotic systems provide the prototype for explor-
ing the complexity of our other, non-material environment. The methodology
for the first of these enterprises derives from physics. For the second, since lan-
guage is the most complex semiotic system we know, probably the methodology
will have to come from linguistics, with grammatics (the theory of grammar)
in the role of mathematics. But in that case I think we need first to dispel some
misconceptions about the nature of meaning in language.

V
There are three rather different assumptions that may be made about meaning
in relation to language: different although also, I think, interrelated. The first
is, that meaning is restricted to ideational meaning: it is ‘content’. The second
is that meaning is a form of representation: it is ‘symbol’. The third is that all
meaning can be measured: it is ‘information’. Let me try to say a word or two
about each.
The first assumption appears most extremely in the metaphor that is
entrenched in the term content; this is related to – indeed is a component of
– the ‘conduit metaphor’ whose pervasive effects were discussed in a famous
article by Reddy (1979). If we discard that metaphor in favour of, say, meaning
as reference, we need not deny that the realm of meaning may make reference
to the realm of matter, provided we recognise that there is often no analogue
70 linguistics and the human sciences

in the real world: much science discourse, for example, is about virtual entities
which exist only on the semiotic plane. The term ideational meaning is intended
to overcome these constraints; and in doing so it helps to point up what is the
real issue here, namely that meaning is not limited to construing. Meaning is
also enacting. If I frown at you, or make a threatening gesture, we can of course
describe these actions in representational terms – we can name them, as I
have just done. But in themselves they are acts of meaning of a different kind,
where meaning is not a mode of reflection but a mode of action. I can mean
at you, so to speak. The problem with terms such as ‘emotional content’ and
‘self-expression’ is that the act of talking about our attitudes and our emotions
transforms them from enactment into construal. But their mode of meaning is
as semiotic process, not as content; this becomes clear when the emotions are
verbalised, as in swearing at somebody, where the meaning is obviously not the
experiential content of the wording. In the terms of our functional grammatics,
meaning is not only ideational, it is also interpersonal.
This is thus related to the second of the assumptions I referred to: that meaning
can be understood as symbolic representation. We are accustomed to thinking
of meaning in terms of symbols, of something that is ‘standing for’ something
else, like the letter of an alphabet standing for a phoneme of the language;
writing probably provided the unconscious model for this way of thinking. But
in that case we have to locate that ‘something else’, the entity that must exist
before anything can stand for it. As Lamb has pointed out, this becomes highly
problematic when we try to account for meaning in neurocognitive terms.
A semiotic system is best thought of not as a set of symbols but as a system
of connections. Sydney Lamb has devoted a great deal of time and energy to
developing and working out such a connectivist model, explaining language as
relational networks in the human brain. The network representations of sys-
temic functional linguistics, which are not of this kind (they are analytic tools,
not refractions of the brain), are quite compatible with Lamb’s account, which
embodies the paradigmatic dimension from Saussure and Hjelmslev instead of
the flat, syntagmatic picture inherited from structuralism. Lamb’s model will
serve as prototype for semiotic systems in general (Lamb, 1999; 2004).
The third assumption was one I questioned earlier, namely that meaning can
be measured in bytes – that it is just another name for ‘information’. I am not
saying that no aspect of the meaning potential of language can be quantified.
We can represent at least some of it in an assembly of system networks, which
are networks of semiotic options. Imagine a network consisting of twenty-four
binary options, all independent of each other and all equally probable; it is easy
to calculate the information contained in such a network. Linguistic systems
are not always binary; they do not operate independently – there is generally
some partial association among them; and their terms are not always equally
 M. A. K. Halliday 71

probable. Moreover there are many thousands of them at the stratum of gram-
mar alone. But in principle it is possible to measure the meaning potential of
a linguistic network, in terms of information and redundancy as defined in
information theory (Shannon & Weaver, 1949; Halliday & James, 1993). At
present we are far from having the data that would enable us to do this. But even
if we did, this would not exhaust the human potential for meaning. Systems
of values, moral, aesthetic and so on, while they certainly constitute types of
semiotic system, cannot be reduced to bytes of information. Like emotions, they
can be glossed and discussed in language, in commonsense terms, and we can
also construct abstract theories about them. But phenomena of any kind can
be discussed, and theorised; and even if a scientific theory, which is a designed
form of semiotic, can itself be quantified as information, that tells us nothing
about the phenomenon which the theory is designed to explain. (It is by no
means certain that the meaning contained in a theory is quantifiable anyway.
Physicists like to maintain that their theories are, or should be, elegant; and I
doubt whether elegance can readily be measured in bytes.)
Of course there are semiotic systems that display some or all of these con-
straining features. One example would be the original system of lexigrams
devised by Duane Rumbaugh for investigating the semiotic capabilities of
bonobo apes, where each of a limited number of visual signs, all equally at risk,
stood for a particular entity in the apes’ environment (Benson et al., 2002). But
it would be a mistake, I think, to consider them as characteristic of the semiotic
realm as a whole.

VI
About 25 years ago we started being told that we were living in the ‘Information
Age’. This was the new term in the series ‘Stone Age; Bronze Age; Iron Age;
Machine Age’. It indicated yet another technical advance, but this time one of a
rather more fundamental kind. If, as the name implied, the exchange of goods-
and-services had been overtaken by the exchange of information as the primary
mode of human interaction, this meant a massive shift away from the realm
of matter towards the realm of meaning, at least for some (small, but power-
ful) minority of the human race. New problems were arising, and hence new
strategies for dealing with them: storage and management of data, information
protection and encryption, information overload and how to survive it – the
semiotic analogue of surviving through a landslide or an avalanche. It was not
long before information technology and information science found a place in
research activities and in the curricula of educational institutions.
The information machine was of course the computer, which extended human
semiotic power in the way that tools and machinery had extended our material
72 linguistics and the human sciences

power. Like all machines, the computer is created by the human brain, as are
the meanings which it processes and the meanings by which it is controlled. But
we have as yet no clear concept of semiotic energy, covering brain power and
computing power. This disjunction may be one reason why, as it turns out, we
are not very good at exchanging information. Industrial accidents like those at
nuclear power stations seem often to be caused by straightforward information
failure: a critical message has not been delivered to where it was needed. Some
five to ten generations ago my own forebears came out of the English coun-
tryside to work the new machinery, in mines and factories and shipyards; and
they learned the new techniques extraordinarily quickly. Now their descendants
have had to learn further skills, manipulating meaning instead of matter; they
have no problem with the computer itself, which is just another machine, but
find it harder to manage and evaluate its output. Information quality control is
on the whole not notably successful5.
Robert Logan, following up the insights of Marshall McLuhan from some
40-odd years ago (Logan, 2000; McLuhan, 1962; 1967), identifies six stages
in the evolution of human semiosis: speech, writing, mathematics, science,
computing, and the Internet. He refers to these as ‘modes of language’, with
the Internet figuring as ‘the sixth language’ in the title of the book. I would
see them rather as semiotic modes, or modes of meaning, since their systemic
relation to language is quite variable: writing is an extension of the functions of
language, mathematics is an abstract tool deriving from language, and science
is a form of knowledge construed in language; but that is not to reject Logan’s
central thesis, which is that there has been a continuous process of the extension
and elaboration of semiotic power throughout the known chapters of human
history. Each new phase makes new demands on human capabilities and human
institutions (see Logan, 2000: Chapter 7); and also brings new maladies, as each
technical advance (in McLuhan’s phrase) ‘numbs human awareness’: Logan cites
‘information-age maladies’ of ‘future shock, information overload, political
gridlock, resurgent nationalism, and ethnic xenophobia’ (2000: 208).
How much of these aspects of the late twentieth century’s ‘descent into barba-
rism’ (Hobsbawm, 1997: Chapter 20) can be blamed specifically on information
technology it is impossible to say; I imagine Logan would agree that, insofar as
technology is implicated as a cause, it is the interaction between the material
and the semiotic components of technology that is critical. For sure they both
equally numb human awarenesss. What the information technology has done
is that it has vastly increased the relative significance of semiotic power in
the total organisation and operation of the power structures in society. This
was a central concern of Basil Bernstein in his profound theoretical studies
in sociology and education. Bernstein distinguished control over physical
 M. A. K. Halliday 73

resources (the field of production) from control over discursive resources (the
field of symbolic control):
Agents of symbolic control could be said to control discursive codes, whereas
agents of production (circulation and exchange) dominate production codes
…. I was distinguishing between a complex division of labour of symbolic
control and a complex division of labour in the economic field. Both divi-
sions and their complexities were the products of new technologies of the
twentieth century relayed by the educational system. (Bernstein, 2000: 110)
Elsewhere he comments that his conception of the pedagogisation of knowledge
is ‘part of a more general theory of symbolic control’ (2000: 189). Power, in the
information age, must always include semiotic power: control not just of the
organs that disseminate the discourse (the mass media) but also of the discourse
itself, the meanings that are engendered and exchanged. Making reference to
Bernstein’ s work, Hasan (2004) has suggested that the only form of resistance to
this semiotic aggression lies in developing a different kind of literacy: a ‘reflec-
tion literacy’ which provides the receivers with the metasemiotic resources to
remove the wool that is all the time being pulled over their eyes. Otherwise, the
democratic forces of the Internet may not be enough to prevent ‘information’
turning into a semiotic weapon of mass destruction.

VII
At about the same time – 25 years ago – it became fashionable amongst linguists
to replace the term semantic, which means ‘meaning as made in language,
linguistic meaning’ (as distinct from semiotic, which means meaning of all
kinds), by the term cognitive, which had an interdisciplinary flavour and was
more likely to bring in research grants. Since mainstream formal linguistics had
either excluded meaning altogether or reduced it to a secondary commentary
on syntax, cognitive linguistics came to be used to signal that meaning was being
brought back on to centre stage. This was a pity, as Lamb had been using the
term since 1971 as that which ‘attempts to model the information system that is
present in the mind of the typical… individual’ (Lamb, 2004: 418) – as what he
later (e.g. 1999) named ‘neurocognitive’ in order to make his own earlier usage
clear. But it was a pity in another sense also, in that it denied the presence of a
semantic stratum in the organisation of language itself.
In other words, this construct reinforces the structuralist view that language
stops at the wording; meaning is brought in (as ‘cognition’) in order to explain
the wording, but it is located in some other quarter of the mind. This not only
adds unnecessary complication to an overall model of language, as Matthiessen
and I pointed out (1999), but it fails to distinguish between meaning as con-
74 linguistics and the human sciences

strued in language and meaning as a property of other semiotic systems, making

it harder to conceptualise the semiotic as a whole.
Semiotic systems are not like physical systems. Those of us who are not
scientists or philosophers are accustomed to leaving it to the natural scientists
to explore the essential unity of human knowledge: to scholars like Edward
Wilson, whose book Consilience (1998) is a powerful exposition of this theme.
Much of what Wilson says I accept; but he is inclined to suppose that the physi-
cal interpretation of phenomena, so successful in giving us a deep insight into
the material world, provides the model for the whole of human understanding.
To the extent that systems of all kinds are grounded in material substance,
this is unproblematic: meaning is an activity of the brain, and the brain can
be understood in terms of physical and biochemical processes. The ‘mind’ can
be understood, with Susan Greenfield (2000), as ‘personalised brain’, the brain
as it develops in the context of each individual’s unique pattern of experience.
At the same time, we do inhabit a plane of reality other than the material; and
this is a different story.
The rest of us – even grammarians! – should also claim the right (or rather,
accept the duty) to survey the general field of human knowledge, to bring
a perspective which is complementary to that of a physicist, a chemist or a
biologist. There is a value in thinking semiotically about phenomena of every
kind, and so ‘thickening’, adding new dimensions to, our understanding. This
complementary perspective is that of the study of meaning (of ‘semiotics’ in
its original, pre-postmodern sense). Within this non-material, semiotic realm,
the prototypical phenomenon is language. The powerhouse of a language is its
grammar (more accurately, its lexicogrammar, the unified stratum of syntax
and vocabulary); so thinking about meaning means thinking grammatically.
As I put it earlier, grammatics – the theory of grammar – is the mathematics
of a general theory of meaning.
Although semiotic systems are not like physical systems, meaning does
involve an expenditure of energy; but it is energy of a different kind, namely
grammatical energy. For example, there is extra grammatical energy involved
in creating metaphor, which is a cross-coupling between two strata, or levels of
organisation in language, the semantics and the lexicogrammar. Metaphor is an
essential ingredient in much of our highly-valued discourse: poets depend on it,
as we know – but so do scientists, though theirs is metaphor of a different kind,
metaphors that are made with structures and word classes rather than with
individual words (Halliday, [1998] 2004). Poets and scientists present (at least
so we assume) different constructions of reality; but their semogenic strategies
are related, and both take up considerable grammatical energy.
 M. A. K. Halliday 75

VIII
Meanwhile however a different motif is emerging, that of information as the
ultimate foundation of matter. This is associated particularly with the work
of the great physicist John Wheeler; von Baeyer quotes the last of five RBQs
(Really Big Questions) that Wheeler identified, and which were brought up at
his 90th birthday symposium: ‘The suggestion is that the material world – the
IT – is wholly or in part constructed from information – the BIT’ (von Baeyer,
2003: xi). Information has been gradually taking over in the discourse of physics
for many decades, as a necessary element in quantum thinking; as an outsider,
trying to understand what I read (for example, David Layzer’s Cosmogenesis,
1990), I had come to assume that the elementary quantum particles must
belong to a moment in time before matter and meaning were differentiated
(rather like the moment in the life of a new-born baby, whose first actions are
likewise undifferentiated).
This ‘information’, first expounded by Claude Shannon half a century ago
(Shannon & Weaver, 1949), is clearly separated from ‘meaning’: information is
measurable, whereas meaning is not. With reference to the fourth RBQ, ‘What
makes meaning?’ von Baeyer comments (2003)6:
The fourth question, ‘What makes meaning?’ refers to the thorny philo-
sophical problem of defining the concept of meaning. At the same time it
recalls the frustration of engineers who have at their disposal a variety of
methods for measuring the amount of information in a message, but none to
deal with its meaning.
They appear to come together in the concept of ‘total information’, which von
Baeyer introduces in the final chapter, ascribed to Zeilinger and Brukner (von
Baeyer, 2003: 232–3):
‘Total information’ turns out to be the quantum-mechanical implementa-
tion of ‘information content’ … Philosophers call it ‘semantic information
content’, because it depends on the actual meaning of the message, which
Shannon so pointedly ignored. The connection with meaning is made via the
element of surprise.
And ‘surprise’ then turns out to be measurable, still in bits; the ‘bit’ being
redefined as ‘the fundamental quantum of human knowledge’ (2003: 234).
At that point I shall stop: since I have only just got hold of von Baeyer’s book,
and read it only spottily, I have hardly begun to access its semantic information
content. As a grammarian, I am accustomed to thinking in complementarities:
the quantum theory’s ‘realisation that the blending of mutually contradictory
attributes is the normal state of affairs among atomic systems’ (von Baeyer,
76 linguistics and the human sciences

2003: 183) is a useful strategy for understanding grammatical systems also.

Even ‘the square root of not’ (2003: 189) might be helpful, when we analyse
how polarity is construed in everyday language. If matter is not just organised
by information but actually made out of information, meaning becomes the
unmarked member of the pair: matter is a special case of meaning, rather than
(as in our usual way of thinking) the other way round. And that is a satisfying
conclusion for a grammarian.

IX
But let me return finally to happenings on a superatomic, human scale. My
concern has been with the two phenomenal realms which we inhabit, and with
the nature of the interpenetration between them. I took the concept of ‘history’
as a point of entry. Given any recognised, or recognisable, historical scenario,
what is the relationship between the events and conditions on the material
plane, on the one hand, and those on the semiotic plane, on the other? Here is
a passage from Christopher Hill’s writings about the English revolution (Hill,
1975: 383–4):
Again and again in this book we have noticed the seventeenth-century radi-
cals shooting ahead of the technical possibilities of their age. Later Biblical
scholarship and anthropology make better sense than they could of the
mythological approach to the Bible; cheap and easily available contraceptive
devices make better sense of free love. Modern physics and chemistry are
catching up with the dialectical element in their thought; modern anthropol-
ogy is a science of society which does not rely on the stars, modern theories
of painless childbirth make no theological assumptions about the Fall of
Man. The concept of evolution makes it possible to conceive of a universe
with no external first cause. The technological possibilities may now exist
even for a community in which the creation of unemployment need not be
regarded as a principal task of government, and in which ‘the beauty of the
commonwealth’ could take precedence over private profit, national power or
even the G.N.P. My object is not to patronize the radicals by patting them on
the head as ‘in advance of their time’ – that tired cliché of the lazy historian.
In some ways they are in advance of ours. But their insights, their poetic
insights, are what seem to me to make them worth studying today.
We cannot, unfortunately, listen in to the conversations of those seventeenth-
century radicals, nor even directly to their political speeches and their sermons.
But at least some portion of their meaning is accessible to us, through the word-
ings that were written down and survived for future generations. Discourses of
 M. A. K. Halliday 77

this kind make it possible to write the history of meaning, as it matches – and
mismatches – what is known about the history of doing.
A history of meaning has both a qualitative and a quantitative aspect.
Qualitatively, there will be certain key discourses which carry especial value,
either intrinsically, because they somehow distil the semiotic essence of their
moment in space-time, or extrinsically because they played a critical part in
the ongoing material events (Martin, 1999; Butt, Lukin & Matthiessen, 2003).
Quantitatively, on the other hand, dominant semiotic motifs emerge more or
less gradually over time; to access and evaluate these one needs a corpus of
contextualised discourses that can be examined and interpreted as a whole (see
e.g. Fairclough, 2000). (On the conjunction between qualitative –‘text-based’
– and quantitative –‘corpus-based’ – investigation of discourse, see Thompson
& Hunston, 2004.)
The title of this new journal raises the issue of linguistics in relation to the
human sciences, and it seems to me that linguistics faces three demands in such
a context. It has to serve as means, as model, and as metaphor. It is a means, to
the extent that it provides a theory of language by reference to which instances
of discourse can be explained and evaluated. We usually do not need linguistic
theory in order to make sense of the text, provided it is in a language that we
know. I have taken history as my ‘text’, so let me borrow one of Christopher
Hill’s seventeenth-century examples (Hill, 1969: 135): when we read the popular
verse:
The law locks up the man or woman
That steals the goose from off the common;
But leaves the greater villain loose
Who steals the common from the goose.

we have no difficulty either in conning it or in relating it along various dimen-

sions to its historical background, both semiotic and material. But there are many
levels of meaning in language, and these may call for specialised techniques and
strategies; for example, Matthiessen’s ‘text profiling’ showing semantic biases in
the selection of options across various text types, or Hasan’s semantic networks
bringing out the different ‘coding orientations’ of families with different social
profiles (Hasan, 1992a; 1992b; Matthiessen, 1999; 2004). Martin and Wodak
(eds) (2003) presents detailed exemplars of linguistic approaches to meaning in
historical discourse; see especially Martin’s chapter ‘Making history: grammar
for interpretation’7.
Linguistics needs also to provide a model. Language can be seen as the
prototype of a semiotic system; hence a theory which is designed to represent
the multidimensional ‘architecture’ of language (Matthiessen, 2003) should be
‘thick’ enough, and rich enough, to offer insight into other semiotic systems.
78 linguistics and the human sciences

Not that all vectors of language are going to be found in the same interrelations
in all other such systems – they clearly are not; but a theoretical account of
language would serve as a conceptual tool with which to explore other forms of
human semiosis, for example images (Kress & van Leeuwen, 1996), the visual
arts (O’Toole, 1994) and music (McDonald, forthcoming). Verbal art presents
a special challenge, since, as noted earlier, it is language operating with distinct
value systems and distinct levels of interpretation (Hasan, [1985] 1989; Butt,
1984; 1988; Thibault, 1991).
Every theory is a metaphor for what it is theorising; linguistics, as has often
been pointed out, has the special property that, whereas in other domains the
theory consists of language but the object does not, in the case of linguistics
both theory and object consist of language. Hence linguistics shares many of
the properties that characterise language itself (cf. Halliday, [1997] 2003). In
more general terms, language theorises the human condition by providing a
semiotic trajectory between the human organism and its ecosocial environ-
ment: between the environment of material and social processes at one end
and the meaning-making organs of speech and hearing at the other (Thibault,
2004). Linguistics, in its turn, theorises that trajectory: it construes meanings
to explain the processes of meaning. Thus as well as providing a metaphor
for language, at another level linguistics stands as metaphor for the whole of
meaning as theorising – for the ability of the semiotic realm to construe itself
into successive planes of virtual reality, in the (so far) unremitting human effort
to understand.
Whatever is not matter, is meaning. All the phenomena of human conscious-
ness – those that are construed by the grammar as processes of sensing (in
English, verbs like think, believe, know, like, hope, want, fear, see, hear and their
nouns thought, idea, belief, knowledge, love and so on, as well as the general term
mean(ing) itself) – belong to the non-material realm that we inhabit, for which
the unifying concept is that of semiosis. Like processes of saying, to which they
are closely agnate, they can project: they contain a semantic feature ‘mean’,
and they are treated by the grammar as semogenic – they create meaning,
and ascribe it to a source, as in we all believe the sun will rise tomorrow. Part
of our experience is the experience of meaning; each language has its way of
transforming the experience of meaning into meaning.
It seems that meaning was how the world began, as quanta (bits, or ‘qubits’)
of information; and that these remain the ultimate constituents of matter. As
matter leads on to life, and then blood, and warm blood, and brains, meaning
turns up again at a higher level, this time as qualia rather than (just) quanta. The
science of meaning of this kind – meaning in the biological, and eventually the
human, sphere – is semiotics; since language is the leading edge of meaning, the
leading edge of semiotics is linguistics. If information is indeed the basic stuff of
the universe, then physics will turn out to be one kind of linguistics, after all.
 M. A. K. Halliday 79

Notes
1 Cf. ibid. p. 168: ‘There can be no legitimate doubt that in the course of the past 200
years the material conditions of the population in the ‘advanced’ countries of the
world have, on average, substantially improved. The fact cannot be seriously disputed
… and insofar as there are historical theories resting on the assumption that it has
not taken place, such theories are wrong’.
2 The actual situation is more complicated – and more interesting – than this simple for-
mulation allows for. It is true that the language has no critical effect (though it might
influence the order in which phenomena come to be investigated). But the script may
have an effect that is more long-lasting. When the Phoenicians took over the Egyptian
charactery, they found it unsuitable for their (Semitic) language, and developed it into
a syllabary; and when the Greeks took over that, they found it unsuitable for their
(IndoEuropean) language and developed it into an alphabet. Now, a letter is patently
a symbol; so the idea of using it to stand for something else, such as a number, or
a formal relation, comes rather easily. But a character is not obviously a symbol; it
shouts its meaning at you, and the Chinese (for whose language a charactery – a mor-
phemic script – is very well suited) never did detach the character from its meaning
(its morpheme) and use it as a purely abstract symbol. Thus both their mathematics
and their linguistics evolved without the use of symbols (Tong, 1999; Halliday, 1981).
Even when the characters were used to transcribe other languages, e.g. in the Sino-
Mongolian script of CC13–14, they were never treated as purely arbitrary phonetic
symbols; the number of characters in Sino-Mongolian writing was about twice
the number of distinct syllables represented. So it is possible that the way Chinese
science evolved, with very little use of symbolisation, was related to the nature of
Chinese writing; likewise the fact that the Greeks, the Indians and later the Arabs did
introduce symbols into their forms of organised knowledge was related to the fact that
their writing systems already depended on the arbitrary nature of the written symbol.
3 More accurately, the second is the ideational, which includes a logical component that
is in complementary relation with the experiential, the two together constituting the
resources of ‘construal’, the construction of reality. (Halliday & Matthiessen, 1999)
4 And meanings that are ‘worded’ may not be externalised; they may remain as
silent verbalisations. These are not too difficult to access if required (as when one is
composing a written text). Much more ineffable are the thoughts ‘that lie too deep for
words’ – but these too are unambiguously processes of meaning.
5 The problem with the exchange of information seems to be that both the process of
exchange and the commodity being exchanged are semiotic; whereas in the exchange
of goods-and-services, the process is semiotic but the commodity is material. I
noted this first when studying children’s language development: goods-and-services
are exchanged very early, in the protolanguage phase (in infancy), but exchanging
information – telling someone of an experience that you know they have not shared
with you – is something that develops only some way in to the learning of the mother
tongue. (The exchange of goods-and-services need not be accompanied by wording
– though with human beings it usually is; but it does involve a sequence of acts of
meaning.) See Halliday ([1975] 2004).
6 And to me it recalls the frustration of early computer engineers when they discovered
that, while they had a variety of methods for looking up words and morphemes in a
dictionary, they had no success at all with machine translation.
7 Reading the well-informed and scholarly accounts of ‘memories’ of the Second World
War in the European Review (11.4, October 2003), under the heading ‘Focus: history
and memory’, I felt how valuable it would be to be able to undertake a linguistic
analysis of some of the discourses concerned.
80 linguistics and the human sciences

References
Benson, J. D., Fries, P., Greaves, W., Iwamoto, K., Savage-Rumbaugh, S. and Taglialatela,
T. (2002) Confrontation and support in bonobo – human discourse. Functions of
Language 9(1): 1–38.
Bernstein, B. (2000) Pedagogy, Symbolic Control and Identity: theory, research, critique.
Revised edition. Lanham MD: Rowman & Littlefield.
Butt, D. G. (1984) The relationship between theme and lexicogrammar in the poetry of
Wallace Stevens. PhD thesis, Macquarie University.
Butt, D. G. (1988) Ideational meaning and the existential fabric of a poem’. In R. P. Fawcett
and D. J. Young (eds) New Developments in Systemic Linguistics Vol. 2: theory and appli-
cation. London and New York: Pinter.
Butt, D. G., Lukin, A. and Matthiessen, C. M. I. M. (2003) Grammar – the First Covert
Operation of War. Macquarie University: Centre for Language in Social Life,
Department of Linguistics.
Childe, V. G. (1942) What Happened in History. Harmondsworth: Penguin Books.
Crombie, A. (1994) Styles of Scientific Thinking in the European Tradition. London:
Duckworth.
Elvin, M. (1973) The Pattern of the Chinese Past. London: Eyre Methuen.
Fairclough, N. (2000) New Labour, New Language? London: Routledge.
Greenfield, S. (2000) Brain Story (video recording) London: BBC Worldwide.
Halliday, M. A. K. [1975] (2004) Learning How to Mean. London: Edward Arnold.
[Reprinted 2004 in Collected Works of M. A. K. Halliday Vol. 4. London: Continuum.]
Halliday, M. A. K. (1981) The origin and early development of Chinese phonological
theory. In R. E. Asher and E. J. A. Henderson (eds) Towards a History of Phonetics
123–39. Edinburgh: Edinburgh University Press.
Halliday, M. A. K. [1987] (2003) Language and the order of nature. In N. Fabb et al. (eds)
The Linguistics of Writing: arguments between language and literature. Manchester:
Manchester University Press. [Reprinted 2003 in Collected Works of M. A. K. Halliday
Vol. 3. London: Continuum.]
Halliday, M. A. K. [1990] (2003) New ways of meaning: the challenge to applied linguis-
tics. Journal of Applied Linguistics (Greek Applied Linguistics Association) 6. [Reprinted
2003 in Collected Works of M. A. K. Halliday Vol. 3. London: Continuum.]
Halliday, M. A. K. (1993) Analysis of scientific texts in English and Chinese. In M. A. K.
Halliday and J. R. Martin, Writing Science: literacy and discursive power 124–32. London:
Falmer.
Halliday, M. A. K. [1997] (2003) Linguistics as metaphor. In A. Simon-Vandenbergen,
K. Davidse and D. Noël (eds) Reconnecting Language: morphology and syntax in func-
tional perspectives. Amsterdam and Philadelphia: John Benjamins. [Reprinted 2003 in
Collected Works of M. A. K. Halliday Vol. 3. London: Continuum.]
Halliday, M. A. K. [1998] (2004) Things and relations: regrammaticizing experience as
technical knowledge. In J. R. Martin and R. Veel (eds) Reading Science: critical and
functional perspectives on discourses of science. London and New York: Routledge.
[Reprinted 2004 in Collected Works of M. A. K. Halliday Vol. 5. London: Continuum.]
 M. A. K. Halliday 81

Halliday, M. A. K. and James, Z. L. (1993) A quantitative study of polarity and primary

tense in the English finite clause. In J. M. Sinclair, M. Hoey and G. Fox (eds) Techniques
of Description: spoken and written discourse 32–66. London and New York: Routledge.
Halliday, M. A. K. and Matthiessen, C. M. I. M. (1999) Construing Experience through
Meaning: a language-based approach to cognition. London: Continuum.
Hasan, R. [1985] (1989) Linguistics, Language and Verbal Art. Geelong, Vic.: Deakin
University Press. [Reprinted London: Oxford University Press].
Hasan, R. (1992a) Rationality in everyday talk: from process to system. In J. Svartvik (ed.)
Corpus Linguistics: proceedings of Nobel symposium 82. Stockholm 4–8 August 1991.
Berlin: Mouton de Gruyter.
Hasan, R. (1992b) Meaning in sociolinguistic theory. In K. Bolton and H. Kwok (eds)
Sociolinguistics Today: international perspectives. London: Routledge.
Hasan, R. (2003) Globalization, literacy and ideology. World Englishes 22(4): 433–48.
Hasan, R. (2004) Literacy pedagogy and social change: directions from Bernstein’s
sociology. Paper prepared for Third International Basil Bernstein Symposium,
Cambridge, 15–18 July 2004.
Hill, C. (1969) The Century of Revolution 1603–1714. London: Sphere Books.
Hill, C. (1975) The World Turned Upside Down: radical ideas during the English revolution.
London: Penguin Books.
Hobsbawm, E. (1997) On History. London: Weidenfeld & Nicolson.
Kress, G. R. and van Leeuwen, T. (1996) Reading Images: the grammar of visual design.
London: Routledge.
Lamb, S. M. (1999) Pathways of the Brain: the neurocognitive basis of language. Amsterdam
and Philadelphia: John Benjamins.
Lamb, S. M. (2004) Language and Reality. Edited Jonathan Webster. London and New
York: Continuum.
Logan, R. K. (2000) The Sixth Language: learning a living in the internet age. Toronto and
New York: Stoddart.
Martin, J. R. (1993) Life as a noun: arresting the universe in science and humanities. In
M. A. K. Halliday and J. R. Martin, Writing Science: literacy and discursive power
221–67. London: Falmer.
Martin, J. R. (1999) Grace: the logogenesis of freedom. Discourse Studies 1(1): 31–58.
Martin, J. R. and Rose, D. (2003) Working with Discourse: meaning beyond the clause.
London and New York: Continuum.
Martin, J. R. and Wodak, R. (eds) (2003) Re/reading the Past: critical and functional per-
spectives on time and value. Amsterdam and Philadelphia: John Benjamins.
Matthiessen, C. M. I. M. (1993) The object of study in cognitive science in relation to its
construal and enactment in language. Cultural Dynamics 6(1–2).
Matthiessen, C. M. I. M. (1998) Construing processes of consciousness: from the com-
monsense model to the uncommonsense model of cognitive science. In J. R. Martin
and R. Veel (eds) Reading Science: Critical and Functional Perspectives on Discourses of
Science. London and New York: Routledge.
82 linguistics and the human sciences

Matthiessen, C. M. I. M. (1999) The system of TRANSITIVITY: an exploratory study of

text-based profiles. Functions of Language 6(1): 1–52
Matthiessen, C. M. I. M. (2003) Systemic Functional Theory: the ‘architecture’ of grammar.
Macquarie University: Department of Linguistics.
Matthiessen, C. M. I. M. (2004) Frequency profiles of some basic grammatical systems:
an interim report. In G. Thompson and S. Hunston (eds) (2005) System and Corpus:
exploring connections. London: Equinox.
McDonald, E. (forthcoming) Through a glass darkly. Linguistics and the Human Sciences
1(3).
McLuhan, M. (1962) The Gutenberg Galaxy: the making of typographic man. Toronto:
University of Toronto Press.s
Toronto, M. (1967) The Medium is the Massage. New York: Bantam Books.
Needham, J. (1958) The translation of old Chinese scientific and technical texts. In A. H.
Smith (ed.) Aspects of Translation. London: Secker & Warburg.
Needham, J. (1978–1995) The Shorter Science and Civilization in China. Cambridge:
Cambridge University Press.
O’Toole, M. (1994) The Language of Displayed Art. London: Leicester University Press.
Reddy, M. (1979) The conduit metaphor: a case of frame conflict in our language about
language. In A. Ortony (ed.) Metaphor and Thought. Cambridge: Cambridge University
Press.
Rose, S. (2003) The Making of Memory: from molecules to mind. London: Vintage Books.
Shannon, C. E. and Weaver, W. (1949) The Mathematical Theory of Communication.
Urbana IL: University of Illinois Press.
Thibault, P. J. (1991) Social Semiotics as Praxis: text, social meaning making and Nabokov’s
‘Ada’. Minneapolis MN: University of Minnesota Press.
Thibault, P. J. (2004) Brain, Mind and the Signifying Body: an ecosocial semiotic theory.
London and New York: Continuum.
Thompson, G. and Hunston, S. (eds) (2005) System and Corpus: exploring connections.
London: Equinox.
Tong, A. K. K. (1999) The language in Chinese mathematical texts before AD 1300: a case
for logical abstraction. Interchange 30(4): 433–58.
von Baeyer, H. C. (2003) Information: the new language of science. London: Weidenfeld &
Nicolson.
Williams, G. C. (1995) A package of information. In J. Brockman (ed.) The Third Culture
Chapter 1. New York: Touchstone Books.
Wilson, E. 0. (1998) Consilience: the unity of knowledge. London: Little, Brown.