How To Avoid Saying Who Did It

6 How to avoid saying who did it
I have been arguing that social scientists, who write in ways that
transform people into things, tend to write imprecisely. When it comes
to describing human actions, ordinary verbs have the edge on the big
izations and ifications. In the previous chapter, I offered a number of
examples to support this view but in the present chapter, I want to go
further: I want to give technical reasons why this should be the case. In
order to sustain my points, I will have to take a number of twisty turns.
Principally I will be drawing on ideas from linguists to show what sorts of
grammatical moves authors make when they write in unpopulated ways
and then I will need to argue why such moves might be problematic for
social scientists.
This will not be straightforward, because I will need to use and also to
hold at a distance some linguistic izations, such as ‘nominalization’ and
‘passivization’. I was faced with a similar problem in the previous chapter,
when I needed to use and to criticize the concept of ‘reification’. I will be
suggesting that even critical linguists, who see ideological problems with
‘nominalization’ and ‘passivization’, fail to heed their own warnings when
it comes to their own writing. My preference is to use the verb forms – ‘to
nominalize’ and ‘to passivize’ – but this will not always be possible,
especially when I follow the arguments of those who use the nouns.
There might also be times when I could have avoided using the izations
myself, but use them regardless: it can sometimes be hard to climb out of
the pit in which we are accustomed to play.
I will be discussing how social scientists, natural scientists and ideo-
logical writers all tend to use the grammatical constructions of nominali-
zation and passivization, but for different reasons. The ideologists do so
because it is often expedient for them to absent themselves from their own
texts and to delete agents from their descriptions of the world. The natural
scientists derive other benefits from writing in impersonal, depopulated
ways. However, it has been a mistake, in my view, for social scientists to try
to follow the same path as the ideologists and the natural scientists. My
argument will require a deviation into the history of scientific writing, but
115
Downloaded from https://www.cambridge.org/core. University of Sussex Library, on 23 Jul 2018 at 19:51:19, subject to the Cambridge Core
terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1017/CBO9781139208833.006
116 Learn to Write Badly
this will be necessary if I am to argue why social scientists should write

differently from natural scientists.
In making my points, I will be summarizing some arguments that I have
presented elsewhere in greater detail (Billig, 2008b and 2008c). In those
publications, I discussed how critical discourse analysts often use the
same grammatical constructions that they claim to have identified as
being ideologically problematic. Some critical linguists responded to my
arguments and they made their critical points in good spirit (see,
Fairclough, 2008a and 2008b; van Dijk, 2008; see also Martin, 2008).
I do not have space here either to reproduce my own arguments, at least in
full, or to discuss the objections to them.
Whereas natural scientists need to avoid personifying the physical
world, social scientists should try to avoid reifying the social world.
Nevertheless, they can derive conventional benefits from using reified
language. I will be suggesting that, with their noun-based language of
things, social scientists can appear to be technically proficient, but by
using technical nouns they can spare themselves the trouble of being
precise. After my twists through linguistics and the history of scientific
writing, I will be returning, at the end of this chapter, to a familiar theme.
Social scientists, writing in ways that omit the agents of particular actions,
not only resemble ideologists and natural scientists, but their turns of
grammar can also resemble those used by advertisers. The style of con-
temporary writing in the social sciences is not always far removed from
the writing styles of promotion and advertising.
Grammar of ideology
How do authorities use language to exert control over others? This was the
question that a group of radical linguists asked in an important book,
Language and Control. Roger Fowler and his colleagues wanted to discover
whether ideological discourse, especially that used by powerful author-
ities, contained particular linguistic features that might distinguish it from
other sorts of discourses (Fowler et al., 1979). Their project was inspired
by George Orwell and his vision of Newspeak as a language which a
totalitarian government invented in order to control the thoughts of its
subjects. In the appendix to his novel Nineteen Eighty-Four (1949/2008),
Orwell had discussed the principles of Newspeak. Its vocabulary would
diminish, not expand; the authorities would invent new words that would
express the official philosophy; nouns and verbs would be interchange-
able. Fowler and his colleagues believed that Orwell was basically correct
to suppose that authorities could maintain their authority through exploit-
ing particular grammatical features, but they questioned whether Orwell
How to avoid saying who did it 117
had identified the correct features. They also rejected Orwell’s assump-
tion that authorities would have to create a virtually new language to exert
control. They thought that those in power could adapt any language,
including Orwell’s beloved English, to their purposes, by using certain
grammatical formulations and avoiding others.
There were two linguistic constructions in particular that Fowler et al.
concentrated upon: ‘nominalization’, or the creating of nouns out of
verbs, and ‘passivization’, which is the turning of active sentences into
passive ones. Fowler and his colleagues claimed that these two construc-
tions feature heavily in official language and later research has tended to
confirm this. It has also confirmed that discourses, which are rich with
nominalization, also tend to be rich with passivization and vice versa
(Biber, 2007). That would not have surprised Fowler and his colleagues
for they suggested that both constructions can be used for similar pur-
poses. By nominalizing and passivizing, authorities can produce formal
documents which describe actions and present orders as agentless things,
thereby achieving what Fowler et al. called ‘agent-deletion’ (p. 33).
For example, authorities often phrase orders in the passive voice – ‘You
are requested not to walk on the grass’; ‘Students are informed that essays
should be submitted before 17.00’. When authorities use the passive, they
can obscure themselves as the authors of these orders and they present
their commands as if they were objective necessities of the world. It is the
same if authorities express the submitting of essays in terms of a noun or
noun phrase: ‘Essay-submission deadline: 17.00’. Authorities will often
favour these grammatical constructions because they allow ‘the details of
the exercise of the mechanisms of control to be obscured, mystified’
(Fowler et al., 1979, p. 41).
Fowler and his colleagues also believed that Orwell was basically correct
in supposing that the coining of new words was vital to maintaining
control over the thoughts of others. As they wrote in relation to a club,
which they studied, an obstinate individual might resist being manipu-
lated on a personal level, but the same person would find it difficult to
evade the control exercised by the new terminology which those in charge
of running the club were introducing. Thus, the inventing of new words –
or what Fowler et al. called ‘relexicalization’ – could be central to the
exercise of control, especially when it came to nouns being formed from
verbs by means of nominalization. As Fowler et al. wrote, ‘nominalization
facilitates lexicalization’ and the invented nouns can be ‘spotted by their
ending in –ion, -ition, -ation, -ience, -ness, -ment’ (1979, p. 40).
Orwell may not have been correct in supposing that authorities exercise
control best through contracting the vocabulary of a language. After all,
nominalization involves creating new terminology and thereby expanding
the vocabulary. Nevertheless, Fowler and his colleagues believed that

both nominalization and passivization led to one vital form of linguistic
contraction. They write that nominalization is a ‘process of syntactic
reduction’ (p. 41); it contracts meaning because it ‘reduces a whole clause
to its nucleus, the verb, and turns that into a noun’ (p. 39).
Roger Fowler was to illustrate this ‘syntactic reduction’ in his book
Language in the News (1991). He looked at the headlines of newspapers, to
see how the headline writers reported an event in which police had
attacked unarmed protestors. It tended to be the left-wing papers that
used active verbs with subjects in their headlines: ‘Police Attack
Protestors’. Right-wing newspapers, which basically supported the police,
tended to omit the police in their headlines. They would, for example,
describe the event in the passive voice: ‘Protestors Attacked’. Or they
would nominalize the event by using the noun ‘attack’ rather than the
corresponding verb: ‘Attack on Protestors’.
In English, if one uses a verb such as ‘attack’ in the active voice, then
one has to indicate the grammatical subject – namely, the agent who
performs the action. By choosing to use the passive voice or by nominal-
izing the event, the headline writers could avoid identifying who did the
attacking. We can see how nominalizing or passivizing leads to a reduction
of information. In using the passive voice – ‘Protestors Attacked’ – the
writer omits the agent but still uses a verb which puts the action in the past.
Even this information is lost in the nominalized ‘Attack on Protestors’ –
the noun does not indicate when the action might have happened or even
if it occurred at all.
Fowler referred to one of the important general conclusions of
Language and Control, when he wrote that ‘we claimed that nominalization
was, inherently, potentially mystificatory’ (1991, p. 80). He went on to
link nominalization with reification:
If mystification is one potential with nominalization, another is reification. Processes

and qualities assume the status of things: impersonal, inanimate, capable of being
amassed and counted like capital, paraded like possessions (Fowler, 1991, p. 80,
emphasis in original).
In this way, Fowler saw the studies, which he and his colleagues had
conducted, as successfully demonstrating how ideology and reification
worked through the use of language. When speakers and writers nomi-
nalized, the end result was not just the hiding of agents but the turning of
human actions into things. No wonder these linguists viewed nominaliza-
tion with suspicion.
However, there is a problem that Fowler and his colleagues did not
confront: nominalization and passivization, also feature heavily in
academic, especially scientific, discourse (Biber and Conrad, 2009; Biber

and Gray, 2010). This raises the conundrum: why should scientists and
ideologists favour the same sorts of grammatical constructions? If the
former are revealing the structure of the world and the latter are hiding
it, one would expect them to use language in very different ways. As the
discussion of ‘reification’ in the previous chapter suggested, social scien-
tists also like to use big, nominalized nouns. Does this mean that their
writing is potentially ‘mystificatory’ or that nominalization is not inher-
ently problematic?
The key to the answer lies in recognizing that, when it comes to
describing human actions, the constructions, produced through nominal-
izing and passivizing, transmit less information than simple sentences with
active verbs. The right-wing headline writers, who concealed the identity
of the attackers, were not doing anything extraordinary. Any competent
user of English should be able to transform an active sentence into a
passive one. Also, they should be able to transform a verb into a noun or
noun phrase. You just have to understand the rules of grammar and the
customs for turning verbs into nouns, even if it means constructing a new
noun by means of suffixes such as ‘ification’ and ‘ization’.
However, these transformations are generally not reversible, for, while
one can always create passive sentences out of active ones with an object,
one cannot always transform sentences the other way round. From the
statement ‘police attacked protestors’, a competent user of English should
be able to construct the sentence ‘protestors were attacked’. However, if
one is faced with the passive sentence ‘protestors were attacked’, one
cannot generate an active sentence unless supplied with further informa-
tion about the nature of the attackers (for example, if the passive sentence
identifies the agents by specifying ‘protestors were attacked by police’).
The phrase ‘attack on protestors’ provides even less information for trans-
forming it into a sentence with an active verb. There is no indication
whether the attack has taken place, is taking place or might take place in
the future. We, therefore, do not know the tense or modality for our verb.
But it does not matter: we cannot in any case create a sentence with an
active verb because we do not know the grammatical subject.
I might appear to be emphasizing some very elementary points of
grammar, which school children might be expected to learn, but I have a
reason. If academic social scientists tend to favour nouns and passives in
their writing, then they are not merely using grammatical forms that are
useful to those who seek to reify the world: they are using forms which
contain less information than sentences with verbs in the active voice.
This means that we cannot readily transform statements about abstract
things, such as ‘reification’ or ‘mediatization’, into sentences with human
actors as their grammatical subjects. Social scientists, in creating their

izations, will have lost key information on the way. It is then little wonder
that their big nouns appear to be imprecisely linked to human actions. We
will see this later, when we look in detail at the concept of ‘nominalization’
itself. What is strange is that social scientists accord higher status to
grammatical forms that contain less information.
Repeating the problem of nominalization

The work of Fowler and his colleagues has been extremely influential, for
amongst other things, it formed the basis of what has become known as
critical discourse analysis, an approach whose abbreviation ‘CDA’ I dis-
cussed in Chapter 4. Critical discourse analysts have systematically devel-
oped the project that Fowler and his colleagues initiated (see, for example,
Fairclough, 1992; Fairclough, Mulderig and Wodak, 2011; van Dijk,
2003 and 2010; Wodak, 2006; Wodak and Meyer, 2010). By and large,
critical discourse analysts have continued where Fowler and his col-
leagues left off. They still treat ‘nominalization’ and ‘passivization’ with
suspicion, considering these to be among the key means for constructing
ideological messages and for hiding the identity of powerful agents. For
example, one critical discourse analyst writes that ‘syntactic transforma-
tions, particularly those labelled “passivization” and “nominalization”
can be considered ideologically problematic because they may obscure
agency – who did what to whom’ (Schroder, 2002, p. 105).
But look at that sentence: its principal verb is in the passive voice (‘can
be considered’). The sentence accuses ‘nominalization’ and ‘passiviza-
tion’ of obscuring agency, but it identifies no agents. It is as if these
syntactic transformations do things by themselves without human help.
Somewhere along the line, the actions of people have been transformed
(by whom?) into things such as ‘syntactic transformation’, ‘passivization’
and ‘nominalization’. These things have then theoretically become the
agents who carry out the business of obscuring agency. The writer seems
unaware that he is using the very forms of language that he is claiming to
find ideologically problematic.
But maybe it is not so odd, for the signs were there all along in the
writings of Fowler and his colleagues. They were warning against the
dangers of using language in ways that suppressed the agency of
people. Yet, they themselves were using nouns such as ‘passivization’
and ‘nominalization’, rather than taking care to formulate sentences
with verbs such as ‘nominalize’ and ‘passivize’ in the active voice with
human agents as subjects. Even when specifically warning how authorities
maintained control by inventing new words, they did not stick with old
words, using a phrase like ‘inventing new words’, but they used the noun
‘relexicalization’. When they warned of writers deleting agents from their
sentences, they then formulated a new noun phrase – ‘agent-deletion’.
They could use this phrase without specifying who was deleting the
agents. It was the same for the idea that nominalizing and passivizing
sentences would reduce meaning: this became the thing ‘syntactic
reduction’.
What Fowler and his colleagues were doing seems to fit Fowler’s own
glorious description of reification. They were creating technical terminol-
ogy for ‘processes and qualities’ and their nouns and noun phrases
‘assume the status of things’ which could be ‘amassed and counted like
capital, paraded like possessions’. We can note the structure of Fowler’s
own description: it too is devoid of people, as he writes about ‘reification’
as a thing, not an activity. Those who might be doing the counting,
amassing and parading are absent visitors. In the description, the only
actors are processes and qualities that supposedly assume the status of
things.
This is not just a metaphorical turn of phrase which authors are using to
capture dramatically an underlying idea. In more prosaic passages Fowler
and his colleagues also make their abstract things act in various ways. As
we have seen, nominalization is said to ‘facilitate’ relexicalization – one
thing paving the way for another thing without humans in the way.
Nominalization has other talents – the authors write that ‘nominalization
can depersonalize, depopulate’ (1979, p. 43). That statement too is
unpopulated and impersonal: the agent, who performs the acts of deper-
sonalizing and depopulating, is the thing ‘nominalization’. The authors
could have written that writers and speakers depersonalize and depopu-
late by nominalizing, but they do not. In the next sentence, they repeat the
linguistic construction, claiming that ‘it [nominalization] can also drain
the language of actional vitality’ (p. 43). They once more present nomi-
nalization as an active doer, while introducing yet another thing – ‘actional
vitality’.
I think that I have given sufficient examples to show that these creative
writers did not apply their message to their own writing. They warned
against creating nouns from verbs in ways that enable writers to avoid
specifying who does what. Yet, they seemed unaware that they were
using this very type of noun. It is as if they could not stop themselves
amassing and parading their own izations. They may have been reacting
against many academic conventions – and they were certainly rebelling
against the intellectual narrowness of much mainstream work in linguis-
tics – but their prose still bears the customary marks of today’s academic
writing.
With the development of critical discourse analysis, there is now much

more than a small group of rebellious scholars who are studying the
grammatical characteristics of ideology. The critical approach to analy-
sing discourse has become a recognized approach, practised and taught
across the world. It is identified by its own acronym and supported by
specialist journals, postgraduate courses, conferences etc. Students, who
wish to study CDA, have much terminology to acquire, if they want to
publish in the specialist journals. There are books and articles to assist
them in their endeavour.
One such article is by Theo van Leeuwen (2010), in his contribution to
a methods textbook on critical discourse analysis, which is aimed primar-
ily at students. Van Leeuwen is an important figure in critical discourse
analysis and has worked with Gunther Kress, one of the original co-
authors of Language and Control (van Leeuwen and Kress, 2011).
Subsequently, van Leeuwen has imaginatively applied some of the tech-
niques of linguistic analysis to the study of visual materials and to music.
Van Leeuwen’s chapter is of particular interest here, because he deals with
the topic of nominalization and he discusses how, by using this and other
grammatical constructions, writers can omit agents from their descrip-
tions of the world. Van Leeuwen is instructing students how to recognize
these constructions and, just as importantly, he is giving students official
names for such constructions. However, as he does this, he uses the same
constructions, and his prose becomes, as it were, a parade of amassed
official things.
In two subsections, respectively entitled ‘Objectivation and
Descriptivization’ and ‘De-agentialization’, van Leeuwen discusses how
writers turn actions into agentless things (van Leeuwen, 2010, pp. 156–7).
These are not the only big words for thing-like processes in his chapter. He
begins the section on ‘Objectivation and Descriptivization’ with a distinc-
tion: actions, he writes, can be ‘activated’; that is, represented as actions,
or they can be ‘de-activated’ or ‘represented in a static way, as though they
are entities or qualities rather than actions’ (p. 156, emphasis in the
original). Note his wording: van Leeuwen writes in the passive voice and
his sentence does not mention any actors doing the representing. If the
technical terms, which he is introducing, seem to be the passive participles
of verbs (activated and de-activated), then van Leeuwen uses these words
adjectivally in a noun phrase, rather than as verbs, when he refers to
‘de-activated representations’.
Van Leeuwen’s focus is upon the de-activated representations, rather
than on the activated ones. He describes how de-activated representations
might be ‘objectivated or descriptivized’. Again this is an agentless passive.
However, van Leeuwen soon moves from verb to noun: ‘In the case of
objectivation, actions or reactions are represented as . . . things, for

instance by means of nominalization or process nouns’ (p. 156). Again
he has used an agentless passive sentence. He then introduces two more
process nouns to indicate two further types of ‘objectivation’ – namely
temporalization and spatialization (p. 156).
According to van Leeuwen, all these are forms of ‘de-agentialization’,
which represent human actions as if they are not actions but are things.
Van Leeuwen does not then give examples of the way that writers or
speakers ‘de-agentialize’ or ‘agentialize’ actions. As a writer he has
moved from verb to noun – from de-agentialize to de-agentialization –
and he is not about to shift back again to the verbs. Instead, he mentions
three different types of ‘de-agentialization’, each of which he describes in
agentless, passive sentences, writing of the way an action or reaction ‘is
represented’. These three types are named with more nominalized proc-
ess nouns: eventuation, existentialization and naturalization (p. 157,
emphasis in original). There is more of the same in the next section,
whose name is ‘Generalization and Abstraction’.
Van Leeuwen is scattering theoretical things around him as he goes on
his way. In the course of a single page, he has produced eight theoretical
things (and a ninth, if you count the old friend ‘nominalization’). He does
this in a section in which he is describing how actions are turned into other
things, with the act of ‘de-agentializing’ itself becoming a thing – ‘de-
agentialization’. Van Leeuwen does not connect his own choice of words
and his rhetorical style to the topic about which he is writing, but he uses
his own terminology as if it were entirely natural.
In this essay, van Leeuwen, is not just providing terminology for stu-
dents, but he is constructing a set of fictional things out of a series of
processes. In effect, he is saying to students, you should be able to spot this
sort of thing and that sort of thing: you should be able to distinguish an
existentialization from a temporalization and you will need these words if
you’re going to do your research properly. As such, he is engaging in what
Fowler et al. (1979) called ‘relexicalization’ (itself an example of relexica-
lixing) and which they identified as a particularly effective way of exerting
control over others.
Van Leeuwen is assisting students by equipping them with a guide for
translating the messy things that people actually write (and say) into a
cleaner set of inter-locking theoretical things. The paradox is that his big
nouns seem to refer to the sorts of thing that he is doing by producing
these big words. He is writing like the theorists of ‘reification’, whom
I discussed in the previous chapter and whose own words seemed to turn
the act of reifying into a thing. It is as if such social scientists are always
looking outwards to others, rather than contemplating whether their
arguments contain messages for themselves, particularly about the ways

that they should be writing.
The problem is that by using the big words, the analyst – whether expert
or apprentice – will not be getting near to describing the action of pro-
ducing big words. Instead, analysts seem to transport themselves away
from the bustle of the world, where humans act, speak and write, into a
fictional warehouse where the big words are neatly stacked, packaged and
ready for distribution. It is then the job of analysts to teach students how to
fetch the requisite packages from the warehouse and to deliver them onto
their own pages.
Things and processes

I expect that most social scientists will reject my criticisms about izations
and they will make an obvious rejoinder. They will say that when social
scientists use such terminology they are neither being ideological nor
failing to be self-critical. They are simply being scientific, for social
scientists, like natural scientists, cannot stick to describing particular
phenomena but must formulate theoretical statements. It is no good
saying ‘Look, there’s a fellow, who’s reifying the world, and there’s
another one who’s confidently issuing orders in a formal, agentless man-
ner.’ The social scientist needs to construct theories that say something
about how reification or de-agentialization might work in general. To
create general theories, they require abstract nouns.
There is nothing remarkable about this, for natural scientists do exactly
the same, when they formulate theories of oxidization or calcification.
How could Darwin have written Origin of Species without the concepts of
‘evolution’ or ‘natural selection’? He might have described this or that
species changing, but he required general concepts to pull things together.
So what is wrong with social scientists writing like scientists and using
grammatical forms such as ‘nominalization’ and ‘passivization’? That
surely is how progress is to be made and any recommendations to avoid
such terminology would only send things backward.
Certainly, there is a strong case to be made that natural scientists need
to label processes with nouns and that nominalizing has proved to be
invaluable for this. In saying why, I will draw upon the arguments of
Michael Halliday, the linguist whose work inspired early critical discourse
analysts, including Roger Fowler and his colleagues. I want to outline
some of Halliday’s arguments about the development of science, for his
arguments can help us to understand why there might be crucial differ-
ences between the natural and the social sciences – and, in particular, why
we need to be careful not to treat human actions linguistically as if they are

physical processes. I should point out that many of Halliday’s followers
will not approve of the way that I will be adapting his ideas for my own
purposes.
We have already encountered Michael Halliday, in earlier chapters, as
the inventor of ‘the ideological metafunction’, which I used to illustrate
the sort of technical phrasing that some academics cannot resist using. It
must be said that Halliday has a fondness for devising new technical
terminology but he cannot be blamed for the ways that others might
misuse his concepts. His continual use of neologisms can be hard going
for outsiders; even fellow linguists can find it off-putting (e.g., Leech,
2006). As much as possible, I will try to avoid using Halliday’s terminol-
ogy, even at the risk of simplifying his thoughts.
Halliday argues that the great scientists of the early modern period had
to be grammatical revolutionaries, for science would not otherwise have
developed. In the days before science, according to Halliday, people
tended to refer to things and processes in different ways. Speakers
would use nouns to denote entities such as material things or people –
‘the rock’, ‘Moses’, ‘God’. When it came to describing processes, or
events occurring over time, speakers would use verbal clauses. They
would say: ‘the trees grew tall’, ‘the rock rolled down the mountain’,
‘God created the heavens and the earth’ etc. These ideas could not be
expressed by a single word. Whether the processes were natural events (‘it
rained for forty nights’), or were human actions (‘Moses killed the
Egyptian’) it required a whole clause to describe them.
Thus, in the early days of language, speakers would use one grammat-
ical form for referring to things and another for processes. Halliday
referred to this as the ‘congruent’ pattern, claiming that it was ‘historically
prior’ in the history of language (Halliday, 2006, p. 107). He also notes
that young children, when learning to speak, use this pattern first with
passive verbs and nouns denoting events coming later. One might say that
the congruent pattern appears to be the natural pattern, for it always seems
to come first, regardless of culture or historical epoch, but Halliday resists
using the word ‘natural’, preferring to use the term ‘congruent’.
Having distinguished between the early ways of describing things and
processes, Michael Halliday then makes a bold claim: if language users
had stuck with the original, congruent pattern, then scientific thinking
would have been impossible. The pioneers of modern science discovered
new processes for which they invented new names. Grammatically, this
was far more revolutionary in its impact than devising names for newly
discovered things. Coining names for new species of plants, which had
been found in tropical forests, or for microorganisms, which could only be
seen under powerful microscopes, was not in itself a revolutionary act.

However, the naming of processes was a different story. The scientific
writers of the early modern period used existing grammatical mechanisms
to create new nouns for processes by using the verbal participle or, as
became increasingly common, by turning verbs into nouns by means of
suffixes such as ‘ization’ or ‘ification’. Halliday claims that ‘the device of
nominalising . . . is an essential resource for constructing scientific dis-
course’ (Halliday, 2006, p. 149). Historically it was the physical scientists,
rather than biological scientists, who led the way, and the movement
coincided with the rise of scientists writing in their own vernacular lan-
guages rather than in Latin (Banks, 2003; Pahta and Taavitsainen, 2004).
When Newton described processes in his book Opticks, he tended
initially to use clauses with nouns and verbs. But then he would often
move to a noun, or noun phrase, which would capture the meaning of
that sentence. At one point, Newton was writing about the humours of
the eye decaying in old age. He commented that when this happens ‘light
will not be refracted enough’ (quoted, Halliday, 2006, p. 68). Here
Newton was explaining the idea clausally. Then, Newton quickly pro-
ceeds to sum up the idea of light being refracted by using the single word
‘refraction’, writing of ‘the want of a sufficient Refraction’ in old age
(p. 68). This sort of move occurs again and again in Newton’s scientific
writings.
Joseph Priestley did something similar in Theory of Positive and Negative
Electricity, where he formulated theories about the ‘repulsion’ and ‘equi-
librium’ of particles. Priestley started by writing clausally about particles
being attracted towards, and repelled by, one another. He then moved to
writing about the ‘mutual repulsion’ and ‘mutual attraction’ of particles,
as well as their ‘equilibrium’ (Halliday and Martin, 1993). Again, the
move is from verbs to nouns, and from using more but shorter words to
using fewer but longer ones, although, as Halliday (2006) has pointed out,
nominalized constructions are often not noticeably shorter than more
clausal ones ‘despite a common belief’ to the contrary (p. 156).
By turning verbs and clauses into nouns, Newton, Priestley and other
scientists of the time were creating new types of things, constructing
science, in the words of Halliday and Martin, as ‘an edifice of things’
(1993, p. 17). Particles repelling one another had become ‘mutual repul-
sion’ and, by this grammatical move, the process of repelling ‘has been
reworded to look like an object: repulsion’ (1993, p. 63). Newton and
Priestley were not writing completely like twentieth-century scientists, for
they did not use noun phrases, comprising only nouns. It would take a
while for phrases such as ‘particle repulsion’ and ‘particle attraction’ to
become commonplace (Biber and Gray, 2010).
Natural scientists continue to find nominalizing useful today. Halliday

(2006) offers the example of a modern physicist writing about the way that
the ‘indistinguishability’ of electrons ‘gives rise to’ an extra ‘attractive
force’ between them (p. 73). An abstract quality – indistinguishability –
becomes the cause of a process, which is, in its turn, depicted as another
thing, namely the attractive force. In this way, it is not the physical objects
themselves, namely the electrons, which are the agents of the action but
abstract qualities and processes, which the scientist is depicting as if
things.
As Halliday has stressed, scientists gain a number of advantages from
writing like this. By nominalizing their descriptions of processes, they are
saved the trouble of repeating themselves: instead of endlessly using
clauses that depict particles repelling one another, the author can shorten
this all to one word – repulsion. More than this, the single word can be
used to sum up what has come before and what the scientist has claimed to
have established. Thus, Priestley builds up an argument with the word
‘repulsion’, just as Newton did with ‘refraction’, for they use such nouns
as if they stand for established, discovered things. Both authors, as it were,
can move forward argumentatively, without having, all the time, to refer
back to earlier clausal statements. Importantly, this means that they can
proceed to devise theories about processes, such as ‘refraction’ and
‘repulsion’. Unless they had named these processes as if they were things,
the scientists would have been unable to formulate theories about them.
This language of things was important in another respect because by
writing in this way the early modern scientists could avoid the danger of
unintentionally personifying the physical world. Newton and Priestley did
not wish to depict physical entities as possessing god-like, or human-like,
powers, for they were propounding scientific laws for the physical world. If
Priestley continued to depict particles attracting, or repulsing, one
another, he would have been using verbs that are typically used about
humans. The particles might then seem to resemble humans attracting or
repelling potential lovers; or worse still, they might resemble god-like
beings. By using concepts of ‘attraction’ and ‘repulsion’ Priestley would
minimize the danger of personifying the physical things about which he
was writing. It was as if the thing-like ‘attraction’ or the thing-like ‘repul-
sion’ were causing the observed movements, not the inner powers or
intentions of the individual particles.
Similarly, by using the noun ‘indistinguishability’, the modern scientist
avoids suggesting that the particles might possess the power to distinguish
different entities. In this way, the writer puts the thing ‘indistiguishability’,
not the act of distinguishing (or not distinguishing), at the heart of things.
Similarly, scientists can find it useful to put verbs in the passive voice, for
the writer then does not have to specify that an agent is doing something.
Instead, light is refracted or particles are repulsed, without any agent
appearing to be engaged in carrying out the refracting or repulsing. It is
no surprise, then, that corpus linguists have found that both ‘nominaliza-
tion’ and ‘passivization’ frequently occur in scientific writing (Biber,
1988; Biber and Conrad, 2009).
Scientists might benefit from writing in these ways, but, as Halliday
points out, there are also downsides. Scientists treat their newly named
processes ‘as if they were some kind of abstract entity or thing’ (Halliday,
2006, p. 20). Halliday refers to these scientific entities, created by nom-
inalizing verbs and clauses, as ‘virtual phenomena’ – virtual entities,
virtual processes – which exist solely ‘on the semiotic plane’ (2006,
p. 123). In this regard, they resemble what Hans Vaihinger called, in his
unjustly neglected book The Philosophy of ‘As If’ (1924/2009), the made-
up fictions, which scientists often use in their theoretical thinking and
which can cause problems when scientists treat their fictional things as
real things. The problem, according to Halliday, is that this edifice of
things tends to convey a static view of the world, which is out of line with
modern physical theories that emphasize fluidity, movement and change
rather than thing-like stability.
The issue here is not whether this way of writing has drawbacks for
natural scientists, but whether it does for social scientists. We can see why
natural scientists might wish to avoid personifying the physical world, by
treating things linguistically as if they were persons. Social scientists face
the opposite problem of reifying the social world, by treating humans and
their actions as if they were things. This is the problem with treating
human processes, such as ‘mediatization’, ‘reification’ and ‘nominaliza-
tion’, as if they were things. If social scientists remove people from the
social world, they will run the risk of writing imprecisely, just as natural
scientists will risk producing poor science if they populate the world. This
is why it is important to examine how social scientists use passives and
turn verbs into nouns.
Scientific writing and the passive voice

By using the passive voice, scientific writers can clear the stage of human
actors, permitting the chosen fictional things to be the stars of the show.
Formulaic phrases using ‘it’, followed by a verb in the passive voice, can be
particularly useful: ‘it can be seen that . . .’, ‘it has been demonstrated
that . . .’, ‘it is argued that . . .’, ‘it will be shown that . . .’, and so on
(Hewings and Hewings, 2006). Whoever is doing the showing,
demonstrating or arguing slips into the background. The object of the

showing, demonstrating and the arguing stands there in the rhetorical
spotlight in full objective glory, ready to take the applause.
Modern scientists are particularly likely to use passive verbs in the
results and methods sections of their research papers (Biber and
Conrad, 2009). Scientific writers did not always write in this way. When
Isaac Newton described what he did in his various experimental inves-
tigations, he had no hesitation in using the first person singular. In his
Opticks, Newton described how I held the prism, how I looked through it,
how I observed the image etc. (Halliday, 2006, pp. 145f.). Nowadays, a
scientific writer is unlikely to write quite like this, but will learn to use
agentless passive verbs: an image was observed, the prism was positioned,
chemicals were refined etc.
When it comes to describing results, modern researchers are also
unlikely to be as personal as Newton was. If a research team is publishing
their results – and many scientific papers these days are multi-authored –
they are unlikely to indicate who precisely did what. Team leaders will not
use their positions of command to write: ‘My statistician/research assis-
tant/postgraduate student found significant results.’ The person, who ran
the statistical tests, will remain rhetorically absent: ‘Significant results
were found . . .’
This is not just a matter of convention but there is a philosophical gain
to be made from presenting less information. By writing about things
being done – and not about people doing those things – scientists can
present their methods and their findings as being independent of the
identity of the researcher. It does not matter, who injected the rats or
who ran the statistical tests, for the results should have been just the
same. This is conveyed by the passive voice: ‘the rats were injected’, ‘the
substance was oxidized’ or ‘analyses of variance were computed’. The
more that is revealed about those who did the work, the more readers
might take this information as relevant. ‘My nice, friendly research
assistant ran the experiment’ might imply that the results depended on
subjective factors such as niceness and friendliness. Imagine what
could be read into the statement: ‘Our Italian statistician found
significant results’. Are the writers subtly alluding to a national stereo-
type whether Italians might be thought to be good at statistics or
might be bad at them? If there is no allusion to a stereotype, why mention
the statistician’s nationality at all? By using the passive voice,
authors avoid all these problems. They are implying that anyone, who
followed the research procedures and who was suitably trained, would
have found identical results, whatever their personality, appearance or
nationality.
The use of passives remains high within the social sciences although in
recent years the first person singular has been making a comeback, largely
as a result of feminist academics challenging the impersonal, masculine
style of writing. However, Biber and his colleagues have pointed out that,
in numerical terms, this has not had a massive effect on academic texts as
the number of times that academic writers are using the first person
singular is small compared with the overall number of passives and
nominalized nouns that they use (Biber and Conrad, 2009; Biber and
Gray, 2010).
Using passives when describing procedures need not always be prob-
lematic for social scientists. Biber and Gray (2010), in their analysis of
academic writing and its grammar, have a section describing the corpus of
materials, which they selected, and the grammatical features, which they
examined. In the first two paragraphs, they use active sentences with ‘we’
as the subject: ‘We collected texts from three 20-year intervals . . . we
consider these as a single group’ etc. (p. 4). In the fourth paragraph,
they switch to passives, when describing the Longmans Spoken and
Written Corpus, which they were using: ‘The corpora were grammatically
annotated . . . more specialized computer programmes were developed’
(p. 4). The identity of those who did the annotating and who developed
the programs was not relevant so long as the research workers did their
tasks appropriately: and the writers convey this by choosing to put their
verbs in the passive voice. Of course, it might also be the case that the
writers did not know who had conducted these tasks, while still trusting
that the anonymous workers had performed suitably. By using the passive,
the writers can avoid naming, or admitting that they cannot name, the
hired hands.
Many linguists deal with written texts and, since they are not actually
interacting with anybody, it does not matter greatly, when describing their
procedures, if they write: ‘I collected seventeen newspapers’ or ‘seventeen
newspapers were collected’. Who gathered up the newspapers, so that
they landed on the analyst’s desk, is not a matter of great import.
However, things may be complicated when social scientists are directly
dealing with people rather than with bits of paper. Whether interviewing
respondents, tape recording conversations or conducting experiments,
they will be interacting in some way with those whom they are studying. In
these cases, it may not be so suitable to use the grammar of the natural
sciences. My example will come from outside linguistics.
Experimental social psychologists pride themselves on being scientific,
and they typically copy the styles of natural scientists. I will be discussing
their linguistic habits in greater detail in Chapter 8, but for now I want to
consider what social scientists can lose by following the rhetoric of natural
scientists. Here is the methods section from an experimental study which

was investigating the attitudes of white Australians towards aboriginal
Australians. The researchers were aiming to find out whether white
Australians, who had aboriginal Australians as friends, would be more
willing to meet a hypothetical aboriginal Australian than would white
Australians without aboriginal friends. The authors describe their
procedure:
Participants were recruited for a study entitled ‘Being Australian’ and completed
the questionnaire in supervised groups of 10–12, and were compensated for their
time with course credit. They were informed verbally and in writing that their
anonymity was protected. Completed questionnaires were placed in an opaque
drop box, and participants were debriefed and given the opportunity to request a
summary of the results of the study (Barlow, Louis and Hewstone, 2009, p. 394).
Of the eight verbs in these three sentences, seven are in the passive voice:
‘were recruited’, ‘were compensated for’, ‘were informed’, etc. In none of
these seven instances do the authors specify who performed the various
actions. The sole active verb refers to an action taken by the participants:
they ‘completed the questionnaire’. The researchers, by contrast, are
grammatically absent from the methods section. We have to assume that
it is the researchers who recruit the participants, protect their anonymity,
do the debriefing etc. One of the passive verbs is particularly ambiguous.
The writers state that the completed questionnaires ‘were placed in an
opaque drop box’. It could have been the participants or the researchers
who did the placing. Probably it was the former but the syntax, linking
three passive verbs by two ‘ands’, would suggest that the unnamed agents
might well have been the same for all three actions. Here is an instance of
something more general. The rhetorical conventions of scientific writing
are often less precise than standard ways of writing, at least when it comes
to describing actions (see Billig, 2011 for more details).
The authors are writing like physical scientists describing the oxidizing
of substances or the injecting of animals. Whatever might be the skin
colour, nationality, gender etc. of the researcher, it is presumed to be
irrelevant when they oxidize substances or inject animals. However, a
social psychological experiment is very different from one in the natural
sciences. We would have good grounds for suspecting that the identity, or
social persona, of the experimenters might very well affect how the par-
ticipants behaved. If the persons doing the recruiting, protecting, debrief-
ing etc. were aboriginal Australians (or even if they were Australians of
African or Maori descent), then the white Australian participants might
have reacted differently, especially when indicating whether they were
willing or not to meet aboriginal Australians.
The general point is that the researchers were not examining something
entirely separate from the social character of the experimenters. Basically,
social psychologists are interested in social relations and when they study
their topics experimentally, they set up social situations – namely, experi-
ments – that involve social relations. The investigators, as they interact
with the participants, are part of these social relations. As a result, there is
no clear separation between the procedures of investigation and the topic
under investigation. This is so different from the experiments that phys-
icists or chemists typically conduct: what they are examining is theoret-
ically distinct from the person doing the examining.
If the authors of the experimental report had used active voices, when
describing their procedures, they would have had to describe the agents of
the experiment. They could have used terms such as ‘we’, ‘the research-
ers’, ‘our researchers’ etc., or even possibly ‘the white Australian research-
ers’. The latter term would have conveyed most information, but, at the
same time, it would have suggested that this extra information – about
nationality and skin colour – might be relevant to the processes that the
authors were investigating (Billig, 1994). It would have underlined that
the social persona of the experimenters might well have been relevant and
might have affected the experimental findings. By not mentioning the
researchers, the writers can avoid the dilemma of choosing how to
describe the researchers and also they can avoid implying that the persona
of the experimenter might have affected the results. The resulting rhetoric
conveys ‘objectivity’ and ‘scientific procedure’, but it accomplishes this by
being vague.
The writer is following convention by describing the processes of the
research in ways that reify them: it is as if the procedures just happen. The
readers know perfectly well that this cannot have been the case, but there is
a tacit conspiracy between journal editor, writer and readers to pretend
that it is. It might seem strange that social scientists, especially those who
pride themselves on being scientific, should systematically describe what
has happened in uninformative ways. In this matter, to appear more
scientific means giving less information. Researchers have to write as if
they have something to hide.
Nominalization: processes and things

There is nothing intrinsically wrong with words such as ‘nominalization’,
‘passivization’ or ‘de-agentialization’. It is not a crime against nature to
construct these nouns or to utter them out aloud. By using terms such as
globalization, privatization, nationalization, we can discuss broadly
what might be happening in the world and what we think should happen.
We will find writers and speakers using these words in newspapers,

magazines and television debates. I am not seeking to ban these words,
as if I were the controller of Newspeak, attempting to reduce the vocabu-
lary available to speakers. The problem is not the words themselves. The
problems come with the ways that social scientists use them, treating them
with exaggerated respect and acting as if they provide the means to
technical precision, while persistently using such terminology in ways
that leave a gap between the concept and the world.
I want to consider ‘nominalization’ as an example. There is no intrinsic
reason why this word should have vague, jumbled meanings, for at face
value it seems to be a necessary, technical term in the discipline of
linguistics, which, in relation to other social sciences, is a comparatively
technical discipline. But, as I hope to show in this section, linguists use the
word in various different ways. The problem is not peculiar to ‘nomina-
lization’ per se, for social scientists will use other technical concepts just as
imprecisely. In Chapter 8, I will touch on the social psychological concept
of ‘categorization’. In many ways, ‘categorization’ is as baggy as ‘nomina-
lization’. The trouble is that the specialists do not handle their big nouns
with care, but they rush to use them, knocking over verbs in their haste and
barging other parts of speech out of the way. In their rush, they fail to tie
the big words firmly to the grounds of human actions, leave them flapping
loosely, but flamboyantly, in the wind.
As we saw in the previous chapter, before Peter Berger introduced the
concept of ‘reification’, he did not catalogue different ways of reifying, or
specify what people might do in order to be said to be reifying. Similarly,
van Leeuwen does not tie concepts like ‘existentialization’, ‘temporaliza-
tion’ and so on to the actions of those that might be considered to be
existentializing or temporalizing. Again, the nouns come first. But isn’t
this similar to Newton and Priestley, formulating their special technical
terms and saving their readers the trouble of reading and re-reading
long-winded clause after long-winded clause?
There is a crucial difference. The human actions, which social scientists
summarize with their technical terms, are unlike the movements of par-
ticles that Newton and Priestley were describing. Both natural and
social scientists present their technical terminology as if it supersedes,
and compensates for, the deficiencies of ordinary language. However, in
the case of the social sciences, ordinary language is very much part of
the world that is being described. In ‘temporalizing’ or ‘reifying’ or ‘de-
agentializing’, people will typically be acting in ways that involve
uttering far less exalted words. Social scientists generally fail to root
their terminology within this world of actions and ordinary language.
Normally physicists and chemists try to tie their words for physical
processes closely to their own experimental procedures and to the move-

ments of the physical entities that they study. The big problem is that
the ordinary way of talking about human actions – with identifiable agents
and verbs in the active voice – carries more information, both in actual
social life and in the world of theory, than does the apparently scientific
style.
Of course, social scientists, unlike ordinary speakers or writers, typically
propose definitions for their key terminology, as if that prevents their
official words from becoming wayward. Norman Fairclough is one of
the leading critical discourse analysts, having written extensively and
perceptively about how those in power use language ideologically. In his
book Discourse and Social Change (1992) he discusses the issue of nomi-
nalization and specifically offers a definition of the concept: it is ‘the
conversion of a clause into a nominal or noun’ (p. 27). We can note the
phrasing: Fairclough does not reverse, or unpick, the process of nominal-
izing by defining the noun in terms of a clause, which might specify the
action of nominalizing. Instead, he defines the noun in terms of a noun
phrase, and, more specifically in terms of another abstract term ‘conver-
sion’. Thus, Fairclough’s definition of ‘nominalization’ comprises agent-
less terms and does not specify what anyone might have to do, in order to
nominalize or to convert clauses into nouns.
Most critical discourse analysts would broadly accept Fairclough’s
definition that ‘nominalization’ denotes a process by which a clause is
converted into a noun. The question is how the process occurs – or rather,
what people have to do in order to make the process occur. If one looks
closely it becomes clear that linguists are not consistent in their use of
‘nominalization’ but use the word to cover very different sorts of process
(for details, see Billig, 2008b and 2008c). It is as if one word fits all and it
does not matter if there is terminological chaos behind the plausible
definition.
Some linguists use the word ‘nominalization’ to refer to the etymolog-
ical processes by which particular sorts of new nouns, especially those
denoting processes, enter the language. First, someone has to formulate
this new word, deriving it from an existing verb or from a clause; and then
others have to take up the new word so that it becomes an accepted part of
the language’s vocabulary. In this sense, ‘nominalization’ would describe
the sort of events which must have occurred, when towards the end of the
nineteenth century the word ‘nominalization’ entered the English lan-
guage and became established, not in everyday speech, but in the techni-
cal vocabulary of linguists. Accordingly, some linguists use
‘nominalization’ to refer to a historical process which results in words
like ‘nominalization’.
By contrast, other linguists – and sometimes the same linguists at other

times – do not use ‘nominalization’ in this historical sense, but will use it
to describe a process that occurs within a single text. They will describe
how an author moves in the course of a single piece of writing from
describing a process in terms of a clause to ascribing a single word to
name that process. In this form of nominalization, the author does not
have to be the person who invented the noun in the first place. When
Halliday was referring to nominalization in the works of Priestley and
Newton, he was principally using the term in this way. He was not
necessarily implying that Newton was the etymological inventor of ‘refrac-
tion’ or that no one had used the word ‘attraction’ before Priestley. But he
was suggesting that these writers were putting these words to new uses, as
they moved from using clausal descriptions to using these nouns to denote
what they had just described in those clauses. Here ‘nominalization’ does
not refer to a historical process, but a process within a single text.
There are further senses of the term. Some academics have used the
word to refer to the psychological process by which an individual person
might cognitively transform a clause into a single noun. This could
happen, for example, when someone is writing an official set of regula-
tions. They will be thinking of the rules in terms of plain, clausal language,
and then they will convert these thoughts into something more official-
sounding, which they will then write down. Something similar might
occur with ‘passivization’. You might think a thought in the active voice
and then turn it round into the passive. Linguists, who follow transforma-
tional grammars or who believe that we are liable to think first in ‘con-
gruent’ ways, might well use ‘nominalization’ and ‘passivization’ to refer
to the ways that people might grammatically transform their inner mental
thoughts.
Then there are linguists, who use ‘nominalization’ to refer to some sort
of ideal process, rather than one that occurs in time, whether that be
historical, mental or even textual time. Fairclough (2008b) has claimed
that many linguists principally use ‘nominalization’ in this ideal sense.
Languages possess grammatical rules by which nouns can be formulated
from verbs, or passive voices from active ones, etc. ‘Nominalization’ and
‘passivization’ can refer to these grammatical rules of transformation,
rather than to processes by which people actually use the rules. It is, of
course, a moot point how one could know that these rules exist, unless
people perform the action of turning verbs into nouns etc. Anyway, this is
another meaning of ‘nominalization’. When linguists use the word in this
sense, then they are referring to an ideal grammatical process – a con-
version – that does not occur over time as actual processes do. For many
non-linguists, who see things as either happening or not happening, this
sort of a process can appear a bit baffling, or to use Fowler’s term, some-
what mystificatory.
So, we can see that linguists do not agree among themselves that there is
one particular sort of process which should properly be described as
‘nominalization’. Instead they are liable to ascribe the word to very differ-
ent kinds of processes, including a process which is not actually a process.
That, however, is not the limit to the semantic anarchy. Despite the
definitions describing ‘nominalization’ as a process, some linguists use
the word to refer to a linguistic entity rather than a process. Corpus
linguists, like Douglas Biber, will say that they are counting the number
of nominalizations within a particular body of texts; they identify these
nominalizations as nouns with endings such as ‘-tion’, ‘-ness’, ‘-ment’ etc.
When they do this, they are treating ‘nominalizations’ as the linguistic
entities, and not as the processes by which these entities may have been
produced either etymologically or textually.
As corpus linguists count the number of ‘nominalizations’ within col-
lections of materials, so they treat ‘nominalization’ as a count noun – that
is, a noun that takes a plural. Count nouns refer to entities that can be
counted and corpus linguists, in defining ‘nominalization’ as a particular
sort of noun, can count ‘nominalizations’. However, those linguists, who
consider ‘nominalization’ to be a process, treat it as a mass noun, in
common with many other social scientific izations and ifications, which
denote processes. For example, ‘globalization’, ‘massification’, ‘mediati-
zation’ and ‘reification’ are all mass nouns: you cannot have four ‘massi-
fications’ or eight ‘mediatizations’, just as you cannot count ‘salt’,
‘electricity’ or ‘hopefulness’.
Regarding ‘nominalization’, here then is further messiness.
‘Nominalization’ can refer either to very different sorts of processes or to
a sort of linguistic entity. It can be used as a mass noun or as a count noun.
It seems that linguists are not disturbed by this outgrowth of uses and
meanings. They still continue to use the word as if ‘nominalization’ really
and clearly exists as something. The surprise is not that this has happened
but that linguists do not seem bothered by it: there is no ‘crisis’ of
nominalization. Linguists, including critical linguists, carry on using the
term without apparently noticing that it comes with a profusion of mean-
ings. If linguists, as specialists in language, do not seem to notice how a
key technical word can pass from being a mass noun to a count noun and
back again, then how are the rest of us expected to notice?
I have come across one researcher using ‘nominalization’ in a very
different sense. This researcher was examining what difference it made
if school teachers addressed their pupils by name (Hilsdon, 1997). The
author must have felt that it was insufficient to write about teachers calling
their pupils by name. This easily understood and easily described action
had to be dressed in an impressive concept: ‘nominalization’. It did not
seem to bother the author – or the editors of the journal in which the article
appeared – that other academics might be using this word in other ways.
Such diversity of meaning is apparently no drawback when it comes to
favouring long words over simple clauses. In fact, it might even be a bonus
for the author can appear to be linking their specific piece work with other
types of research.
Once again, we can see that putting the big izations into place does not
achieve precision, but now we are in a better position to see why not. So
long as you keep using words like ‘nominalization’, you do not have to
specify exactly what they mean and to what sorts of processes they
refer. Moreover, you can formulate theories linking one ization with
another, as if you are solving problems by slotting your fictional things
together. For example, you might assert that ‘nominalization leads to de-
agentialization’. But as you do so, you will be leaving a large gap between
the world of theory and the world of actions – and no amount of further big
nouns will fill that gap, for each will bring its own gap to the theory.
The basic problem is that our clauses contain more information about
social actions than our formal nouns do. It is so much harder to unpick the
big nouns and to imagine what actions comprise the ‘thing’ that they
denote. However, if we use clauses to describe actions, then we will be
more likely to think about what is going on. Suppose we want to formulate
a general rule and instead of saying that ‘nominalization leads to de-
agentialization’, we phrase this as a general clausal statement in the active
voice: ‘When someone nominalizes, they will be likely to de-agentialize.’ If
we put forward this statement, then it would be easy for someone to ask us
(or for us to ask ourselves): ‘How exactly do you nominalize or de-
agentialize? What sort of actions do you have to take? Does it matter
how you nominalize? Can you give me an example of someone nominal-
izing and de-agentializing?’ If we then try to answer such questions, we
will find ourselves going into the sorts of detail that we can avoid when we
stick with the big nouns, treating them as things.
That should give us pause for thought. Perhaps the real function of these
big nouns is not that by using them we can formulate better theories or
that we can clarify what goes on in the world. It could be that we find
them useful for another reason. When we use these nouns, we do not have
to be clear – we do not have to think hard about what we really mean,
especially when we are writing for others who regularly use these same
words. In our own safe circles, where we all will be exchanging the same
semantic tokens, we can leave the gap between the world and the words as
wide as we want.
Promoting the nouns

There is a further rhetorical turn that those who write either ideologically
or scientifically can find useful. Having emptied their prose of people,
acting as agents, they can refill it with things that act like people. In this
way, writers can depict fictional things, rather than people, as the agents of
the world. Critical discourse analysts have described how ideological
writers make this move. There have been a number of excellent analyses
of economic language, suggesting that right-wing economists often reify
the social world by treating economic concepts, such as ‘market forces’, as
if they were objective entities (e.g., Fairclough, 2003, pp. 143ff.; Muntigl,
2002; Mautner, 2005). More than this, they ascribe motives to these so-
called ‘objective entities’, making them act as if they were human. For
example, we can read about market forces that dictate/demand/forbid, as
if these forces spring to life to perform the human actions of dictating,
demanding and forbidding. Moreover, these strange forces seem to crowd
out actual people and their actions. The result is that the humans, who
have power over others through their trading and owning, become invis-
ible. In their place, abstract economic forces take on the role of quasi-
humans, demanding obeisance from the real humans.
Right-wing supporters of current economic arrangements are not the
only people who make these sorts of rhetorical moves. One study com-
pared the frequency of inanimate entities that appeared as the subject of
active verbs in scientific journals and in popular magazines (Master,
2006). This grammatical pattern occurred much more frequently in
scientific journals where writers of research articles often made abstract
concepts the subject of verbs such as ‘make’, ‘use’ and ‘allow’. Jay Lemke
(1995), in his book Textual Politics, discusses how this move can occur in
scientific reports. The authors linguistically delete human agents as the
authors use ‘agentless passive clause structures’ and nominalize the
descriptions of processes. Then, according to Lemke, the authors present
‘the nominalized processes . . . as agents in the place of human agents’
(p. 60).
In the previous chapter, we saw something similar occur with authors
writing about the concept of ‘reification’. They tended not to depict
people going about reifying, but wrote about what ‘reification’ could do.
It was the same with ‘mediatization’ and psychoanalysts can treat ‘repres-
sion’ in the same way. In this chapter, we have already seen how Fowler
and his colleagues wrote about ‘nominalization’ doing things like depop-
ulating, depersonalizing, facilitating etc. They claimed that nominaliza-
tion ‘permitted habits of concealment’ (1979, p. 80). When analysing
regulations for applying to university, they commented that ‘the
passive structure, allowing agent-deletion, permits a discreet silence

about who if anyone might refuse to admit the applicant’ (1979, p. 41).
‘Allowing’ and ‘permitting’ are activities, which we usually ascribe to
human agents, particularly those with some sort of power. A police officer
might permit a driver to continue along a road or a parent might allow a
child to eat some chocolate. They have to utter words or make gestures in
order to grant such permission. But how can a grammatical structure be
said to permit actions to occur or not occur? The linguists mean: if we
nominalize or use passive verbs, then we need not specify who is doing the
actions. The linguists are expressing this idea, through their own nominal-
izing (i.e., ‘agent-deletion’, ‘the passive structure’, ‘nominalization’ etc.)
and by ascribing the power of permission to such entities. In their own
statements, grammatical constructions, rather than human agents, have
the power of permitting and forbidding.
Of course, one might counter by saying that this is a convenient,
metaphorical way of writing that prevents endless repetition and awkward
clause-making. And no harm is done by writing in this way: no one really
imagines a nominalization actually standing there, like a policeman in
uniform beckoning the traffic to advance. On the other hand, we could say
that this rhetorical trope is particularly useful if you want to promote your
approach or to emphasize the importance of the concepts that form the
basis of your approach. By depicting the concepts as agents in the world,
you will be accentuating their importance, or, to use current slang, bigging
them up.
Social scientists often use this trope when describing their own theo-
retically favourite concepts and things. In discussing the ‘congruent mode
of grammar’, where a verb means a happening and a noun means an
entity, Michael Halliday comments: ‘But if grammar can construe expe-
rience in this way, it can also reconstrue it in other terms’ (2006, p. 14).
Actually, it is humans, not grammar, that construe or reconstrue experi-
ence although they might do so by means of grammar. This is not a
passing figure of speech. On the next page Halliday claims that ‘the
grammar can always turn one class of word into another’ (p. 15) and he
offers an example where ‘the grammar is reconstruing a happening as if it
was a kind of thing’ (p. 15). A few pages earlier, he had commented that
‘grammar transforms experience into meaning’ and ‘that what grammar
does is to construct a semiotic flow – a flow of meaning’ (p. 11).
Seemingly, it is grammar, not people, doing all these things.
Douglas Biber, another linguist whose work I have discussed, compares
academic texts with other genres of writing. He is arguing that academic
writing often is less explicit than other forms of writing, despite the fact
that we assume it to be more precise. He puts his point in the following
way: ‘Academic texts have systematically chosen grammatical styles of

expression that are less explicit than alternative (more explicit) grammat-
ical styles’ (Biber and Gray, 2010, p. 13). Of course, academic texts do not
literally choose their own grammatical styles. Their authors make these
choices, just as Biber and Gray have chosen to make ‘academic texts’ the
subject of their sentence; their text has not done this for them. Similarly
Halliday writes that ‘the text itself creates its own grammar, instantially as
it goes along’ (2006, p. 143). Here one entity, the text, creates another
entity, its grammar, without any people intervening. It is as if there is a
world of things that do our writing and speaking on our behalf: and we, the
experts, are revealing these important hidden actors. Our own destiny is
bound up with the destiny of our creations, for the more important these
thing-like actors are portrayed to be, the more important are our theories
about them.
In earlier chapters, I have given some examples of academics promoting
their own approaches, which they claim to be able to do things. We saw
how supporters claimed that ‘phenomenography’ favours a dynamic
approach and provides good opportunities, ascribing to their approach
the powers of favouring and to providing good things. In the previous
chapter, supporters of ‘mediatization’ talked of their approach in a sim-
ilarly active way. If I am to follow the strategy of taking examples where
I find them, I should provide some examples from the writings of critical
discourse analysts. I do so at the risk of upsetting figures such as Norman
Fairclough and Ruth Wodak, notable scholars whose work I admire.
Because of this – rather than despite it – I am offering several examples
from their writings about critical discourse analysis.
Lilie Chouliaraki and Norman Fairclough begin their book Discourse in
Late Modernity with the statement ‘Critical discourse analysis (henceforth
CDA) has established itself internationally over the past twenty years’
(Chouliaraki and Fairclough, 1999, p. 1). It is as if a thing ‘CDA’ is
capable of establishing itself. Then this thing takes on the character of a
human, capable of doing what humans do: the authors write that CDA
‘can strengthen its analysis of language’ (p. 2), that it ‘can take further’
lines of research (p. 2), and that it ‘brings critical social sciences and
linguistics together . . . setting up a dialogue between them’ (p. 6) and so
on. This thing also has opinions: CDA ‘takes the view that a text can be
understood in different ways’ but ‘CDA does not itself advocate a partic-
ular understanding of a text, though it may advocate a particular explan-
ation’ (p. 67).
Fairclough, Mulderig and Wodak (2011) also ascribe a number of
psychological capabilities to critical discourse analysis. They outline
how CDA ‘sees discourse’ (p. 357) and how it ‘sees itself’ and how it
‘aims’ to make opaque aspects of discourse visible (p. 358). Moreover,

CDA is capable of making claims (p. 368), highlighting various things
(p. 369), and achieving much success (p. 373). Wodak and Meyer (2010)
discuss how it is ‘interested in’ analysing power and that it ‘aims to
investigate’ social inequality (p. 10). And in all its various forms CDA
‘understands itself’ to be based in theory (p. 23).
Of course, one can see the advantage in writing this way. Writers do
not have to keep referring to the analysts, who might include themselves.
They can achieve a measure of distance (and modesty) by talking about
the perspective and ascribing their own aims, intentions and successes
to the perspective. There is no deceit involved. No one really believes that
the approach actually functions like a person, or that it can exist and have
these aims, intentions and successes without people having aims, inten-
tions and successes. Talking about the approach doing all this gives the
approach solidity. It exists as something in itself – something that has
established itself. In this respect, the psychological verbs act like the
acronym: they set the approach in place as an established part of the
academic scene.
This sort of phrasing resembles a style of writing that is common in
advertising. This is a style where the product is presented as if it operates
itself. Marianne Hundt (2007 and 2009) has suggested that advertisers
often use a voice, which is neither active nor passive, but somewhere in
between. She has called this ‘mediopassive’, after the grammatical voice
which linguists have identified in certain ancient Semitic languages.
These new ‘mediopassives’ are slightly different. The verbs do not have
specific grammatical forms, but their grammatical subjects are syntacti-
cally their customary objects. They occur in advertisers’ statements when
the products appear as the grammatical subject, although they are, in
practice, the objects of the actions being described. ‘The curtains are
selling well’: the curtains are not selling anything, but they are being
sold. ‘The car drives smoothly’: someone has to drive the car to find out
whether it is smooth or not. ‘The door opens easily’: the door does not
open itself but it has to be opened. Hundt has traced how this form of
expression has become increasingly common in English over the past
hundred years.
Hundt suggests that ‘mediopassives’ might be used whenever
someone wants to depict a process ‘as independent of external influence’,
or whenever ‘an agent is difficult to identify’ (2007, p. 167). Advertisers,
boasting about the qualities of their products, want to present their
products as performing well, with the good performance not being
dependent on the skills of those who must operate the product.
Accordingly, the advertisers grammatically delete the agent, setting the
product off on its own, running itself, selling itself, driving itself and
establishing itself.
Hundt is at pains to stress that ‘mediopassives’ are not confined to
advertising, but academics also use them. Like advertisers, academics
often wish to describe processes as if occurring on their own. Hundt offers
some examples from her own discipline of linguistics – such as ‘mor-
phemes that attach to roots’ and ‘vowels that raise’ (p. 167). The mor-
phemes do not actually do the attaching or the vowels the raising. The
examples, that I have presented, suggest that academics, especially in the
social sciences, may frequently use this way of writing when describing
their key concepts or their own approach.
We should not automatically assume that advertising and academic
writing are completely different sorts of genre and that it is just by chance
that both advertisers and academics might use similar linguistic construc-
tions for omitting agents. Academics also have products to promote and
they will praise their own theories and approaches, recommending them
to readers. They will want to say that the product operates well – has
insights, produces findings, exposes what is hidden. Anyone using the
product will have these benefits and understandings. It does not take an
unusually skilled individual to reap the rewards, just as any driver would
drive the car smoothly or open the doors with ease. The product seems to
run itself, and readers, particularly student readers, are invited to try it
out.
At this point the mythology of advertising meets the scientific claims
of the social sciences. The result is rather like the mediopassive construc-
tion itself – neither one thing nor the other. This is neither literal nor
metaphorical writing. Social scientists do not offer their constructs as if
they were only metaphors. They offer their ‘reifications’, ‘nominaliza-
tions’ and ‘mediatizations’ as describing real, not metaphorical, pro-
cesses. On the other hand, they do not root these constructs firmly in
the world. Instead these words seem to swirl about, while the writers claim
that the words are doing things that they cannot possibly do. It is the same
with approaches and theories. The more convinced we are of their use-
fulness, the more we write about them in impossible ways. And the more
we commit an error that seems easier to recognize in others than in
ourselves – namely, the error of describing how things happen in the social
world without mentioning how people might make them happen, or,
indeed, who the people are who make them happen.

How To Avoid Saying Who Did It

Uploaded by

Copyright:

Available Formats

How To Avoid Saying Who Did It

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

How To Avoid Saying Who Did It

Uploaded by

Copyright:

Available Formats

6 How to avoid saying who did it

this will be necessary if I am to argue why social scientists should write

the vocabulary. Nevertheless, Fowler and his colleagues believed that

If mystiﬁcation is one potential with nominalization, another is reiﬁcation. Processes

academic, especially scientiﬁc, discourse (Biber and Conrad, 2009; Biber

actors as their grammatical subjects. Social scientists, in creating their

Repeating the problem of nominalization

With the development of critical discourse analysis, there is now much

objectivation, actions or reactions are represented as . . . things, for

arguments contain messages for themselves, particularly about the ways

Things and processes

we need to be careful not to treat human actions linguistically as if they are

seen under powerful microscopes, was not in itself a revolutionary act.

Natural scientists continue to ﬁnd nominalizing useful today. Halliday

Scientiﬁc writing and the passive voice

demonstrating or arguing slips into the background. The object of the

scientists. Here is the methods section from an experimental study which

Nominalization: processes and things

We will ﬁnd writers and speakers using these words in newspapers,

processes closely to their own experimental procedures and to the move-

By contrast, other linguists – and sometimes the same linguists at other

Promoting the nouns

passive structure, allowing agent-deletion, permits a discreet silence

way: ‘Academic texts have systematically chosen grammatical styles of

‘aims’ to make opaque aspects of discourse visible (p. 358). Moreover,

You might also like