The Promise of Plain Language Drafting
Ruth Sullivan*
The purpose of drafting legislative texts in plain language
is to enhance democracy and the role of law by making legisla-
tion accessible to the people whose lives it affects. But plain
language drafting has some potentially more radical implica-
tions, as it reveals and challenges assumptions underlying cur-
rent approaches to statutory interpretation and to law generally.
The author surveys various plain language initiatives,
highlighting problems encountered by the plain
language
drafter as well as techniques for achieving more direct commu-
nication of the legislative message. A central problem for plain
language drafting is identifying the legislature’s primary audi-
ence. The author shows that it is impossible to write for every-
one. After examining different ways of responding to this
problem, she argues that ordinarily the drafter should write for
the most vulnerable group affected by the legislation to be
drafted.
The author concludes that plain language drafting will
not make it easy for members of the public to read and under-
stand legislation. Its chief value lies rather in the message it
sends to the courts and official interpreters of legislation,
namely, that legislation should be interpreted from the perspec-
tive of its primary audience. Instead of relying on judicial notice
of meaning and their own common sense, courts should receive
evidence about the audience for which the legislation was written.
Judicial interpretation should be informed by an understanding of
the context in which the legislation actually operates.
Another virtue of plain language drafting is that it chal-
lenges some of the untenable assumptions underlying a posi-
tivist view of law, in particular, the idea that law exists in ad-
vance of its application because it consists of rules contained in
texts. The techniques used by plain language drafters tend to
detach the law from particular texts and to blur the conventional
distinction between text and context. In this way, plain lan-
guage drafting draws attention to the malleable character of
texts and the make-it-up-as-you-go-along character of law.
La rddaction de textes Idgislatifs en langage simple vise h
protdger la ddmocratie et l’6tat de droit en rendant la lgislation
accessible h ceux qu’elle gouveme. Elie a cependant d’autres
consdquences plus fondamentales, notamment en ce qu’elle
permet d’idontifier et de remettre en question les postulats sons-
jacents aux principales approches de l’interprdtation des lois et
du droit en gdn6ral.
L’auteure analyse divers projets de rdaction en langage
simple et indique les probllmes qui leur sont lids ainsi que cer-
talus moyens permettant de communiquer le message Idgislatif
d’une manitre plus directe. II appert que le probllme essentil
qui se sou~ve dans la rddaction en Iangage simple est
l’idontification du public principal puisque, comme le ddmontre
l’auteure, il n’est pas possible de rddiger pour tons les groupes
sociaux. A la lumiare de l’dtude de diffdrentes solutions A cette
difficultd, l’auteure sugg4r quo le Idgislateur devrait rdiger en
fonction du groupe le plus vulndrable que la loi en question
peut affecter.
L’auteure conclut que Ia rddaction en langage simple ne
rendra pas en soi la lecture et In comprdhension de la ldgislation
simples. Le principal avantage de ce type de rddaction se situe
done au niveau du message communiqu6 aux tribunaux et aux
autres organs devant interprdter la loi : cette demiare devrait
tre interprdtde A partir du point de ,ue de son principal public.
Plutdt que de s’appuyer sur la connaissance judicaire et leur
propre raison, les juges devraient accepter la preuve du type de
loi dtait destinde, de maniare A ce que
public auquel la
l’interprdtation judiciaire se base sur une comprdhension du vd-
ritable contexte dans lequel Is dgislation agit.
La rddaction en
langage simple prdsente 6galement
l’avantage de remettre en question certains postulats de la con-
ception positiviste du droit, notamment l’idde selon laquelle le
droit prdexiste son application puisqu’il est constitu6 de normes
contenues dana des textes. Les techniques utilisdes pour la r-
daction en langage simple sdparent le droit de textes particuliers
et estompent la distinction classique entre le texte et le contexte.
De cotte fagon, la redaction en langage simple met en dvidence
le caraetae malldable des textes et la nature fonciarement ad
hoc du droit.
. Professor, Faculty of Law, University of Ottawa. The research for this paper was funded by a grant
from the Social Sciences and Humanities Research Council of Canada.
McGill Law Journal 2001
Revue de droit de McGill 2001
To be cited as: (2001) 47 McGill L.J. 97
Mode de r6fdrence: (2001) 47 R.D. McGill 97
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 47
Introduction
1. Evolving Theory and Practice
A. Direct Communication
1. Access
2. Understanding
3. Personalization
B. Audience Assessment
C. Reliance on Experts from Other Disciplines
D. User Testing
I1. Writing for Divergent Audiences
A. Writing for Everyone
B. Writing for the Audience Targeted by Parliament
C. Writing for the Least Experienced
D. Writing for Actual Readers
E Conclusion
Ill. The Impact of Plain Language Drafting
A. The Plain Language Message
B. The Death of Positivism?
C. Remaining Issues
Conclusion
20011
R. SULLIVAN – THE PROMISE OF PLAIN LANGUAGE DRAFTING
99
If one, settling a pillow by her head,
Should say: “That is not what I meant, not what I meant at all.
That is not it, at all.”
T.S. Eliot, “The Love Song of J. Alfred Prufrock”‘
Introduction
In nearly all jurisdictions, the role of legislative drafting in the creation and ad-
ministration of law receives scant attention from legal educators and the practising
bar. This neglect is a mistake, in my view. Given current conceptions of democracy
and the rule of law, and given the way those conceptions are put into practice in par-
liamentary democracies like Canada, the work that drafters do warrants careful scru-
tiny.
The primary task of drafters is to embody proposed legislation in a fixed form-
to “in-text” future law, in effect. Once established, the text is tabled in the legislature,
where the elected representatives of the people decide whether to enact it, in whole or
in part, with or without amendment. If this exercise in democracy is to be more than
an empty formality, it is obvious that the representatives who vote for or against a
particular proposal must understand what is being proposed.
Once a text is enacted, it is published in an official publication like the Canada
Gazette. Publication is meant to ensure that members of the public, who are deemed
to know the law, have access to the statute book and can come to know the law in fact
as well as legal fiction. In keeping with the rule of law, it ensures that citizens have a
means to identify their rights and obligations as set out in the enactments that apply to
them. In keeping with democracy, it ensures that citizens are able to understand and
assess the output of their elected representatives.
The rule of law and democracy also require that those who administer the law ap-
ply it equally to everyone, exactly as enacted, so as to prevent discriminatory effects
or unauthorized outcomes. Law that has been appropriately “in-texted” facilitates
equal, accurate application because the rules to be enforced and the powers available
to enforce them are set out in a fixed form, which is accessible to all.
As this brief sketch suggests, democracy and rule of law as currently conceived in
Canada presuppose the capacity of a legislative text to embody the law in a fixed form
and to communicate the same law to everyone-the parliamentarians who will enact
the law, the citizens whose rights and interests will be affected by it, and the officials
‘T.S. Eliot, The Complete Poems and Plays (London: Faber & Faber, 1969) 13 at 16.
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 47
who will enforce it. The challenge of creating texts that fulfil these requirements falls
to the legislative drafter.
Historically, drafters have focused on the first challenge, that of “in-texting” the
law, and have paid little attention to the challenge of communication. The result is a
statute book that declares the law in thousands of carefully written pages that are in-
comprehensible to all but a handful of legal insiders. This insider group includes
judges and lawyers, as well as professionals, such as accountants, who deal with cer-
tain areas of the statute book on a regular basis. It may also include bureaucrats who
administer statute-based schemes, especially those who prepare secondary materials
that attempt to explain “their” legislation to the general public. Insiders become fa-
miliar with portions of the statute book and they may become adept readers of legis-
lation in general. To the rest of the public, however, the statute book is a walled for-
tress, bristling with bizarre weaponry. Who would want to go there?
Much legislation is drafted in long, convoluted sentences and relies on obscure
jargon. But this is the least of the public’s problems. The first challenge faced by
would-be users of legislation is getting hold of an accurate and up-to-date copy of the
legislation, including not only the act, but also regulations and documents incorpo-
rated by reference.2 The second challenge is to locate the parts of the legislation that
are relevant. Users must then read those parts and make sense of them, a challenge
that takes more than an ability to decipher badly written prose. Finally, and most im-
portantly, users must appreciate the import of what they have read in terms of their
personal circumstances and interests.
Given these multiple challenges, it is not surprising that access to law for most
people is impossible without the assistance of lawyers or other professionals. These
professionals not only locate the law and explain it, but also apply it in a way that
benefits their clients to the greatest possible extent. They are both expected and
obliged to use their knowledge and skills to develop interpretations that favour their
client’s position. For those who can afford a professional to look after their needs and
interests, dealing with the statute book is not a problem: the professional acts as in-
termediary between the client and the text. For the rest of the public, however, the
statute book remains an intimidating and impenetrable fortress.
2 Incorporation by reference is used frequently in federal legislation and creates significant access
problems. Often the incorporated materials are in English only, are difficult to track down, and ob-
taining copies is prohibitively expensive. When documents are incorporated as amended from time to
time (as occurs in a so-called “ambulatory” incorporation by reference), the problems are even
greater. Incorporated materials may be revised on a regular basis and the revisions are not necessarily
publicized. There are also interpretation problems. It is unclear, for example, whether incorporated
materials should be interpreted in their own literary and operational context or in the literary and op-
erational context of the legislation.
20011
R. SULLIVAN – THE PROMISE OF PLAIN LANGUAGE DRAFTING
101
Most supporters of plain language drafting find this arrangement unacceptable.
They believe that legislation should speak directly, without the need for intermediar-
ies, to the very people whose lives it affects. This obviously is a worthy ambition, but
it may not be a realistic one. In this paper, I review some recent practices of plain lan-
guage drafting and I attempt to assess the assumptions on which they are based and
the impact they may have on the various audiences for legislation, both professional
and non-professional. I conclude that plain language drafting is an important initiative
that deserves scholarly attention and support, not because it is likely to make law
readily accessible to the general public, but because it reveals the false assumptions
underlying prevailing notions of democracy and the rule of law and attempts to do
something about them.
Part I of this paper is largely descriptive. It looks at the evolving theory of plain
language drafting and mentions a number of initiatives undertaken in Canada and
elsewhere to put the theory into practice. Part I considers some of the difficulties en-
countered by drafters in trying to communicate directly, without the benefit of inter-
mediaries, to widely divergent audiences. It ends by suggesting that drafters should
not try to write for everyone; rather, unless directed otherwise, they should write for
the most vulnerable audience affected by the legislation. Part III considers some im-
plications of plain language drafting for statutory interpretation and legal scholarship.
It suggests that plain language drafting should lead courts and other official interpret-
ers to adopt an audience-based approach to interpretation, and it should lead all of us
to think more creatively about ways to enhance democracy and promote the rule of
law.
I. Evolving Theory and Practice
There was a time when drafting in plain language meant adhering to a list of
rules: keep your sentences short, use simple words, avoid the passive voice. However,
this limited and mechanical approach has been discredited. The distinctive feature of
plain language drafting these days is its pragmatic and empirical approach. Once the
legislature’s specific goals have been identified, the drafter looks for practical means
to achieve them, relying on interdisciplinary research and empirical testing to ensure
that those means are effective.
A. Direct Communication
If drafters want to communicate directly with their audience, so that legislation
can be used without the help of intermediaries, then they themselves have to carry out
the work that intermediaries do. In effect, the drafter must provide the services that le-
gal professionals typically provide: finding the relevant provisions, translating them
into comprehensible language, explaining them, and personalizing them. These serv-
ices must somehow be carried out in the legislation itself.
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol.47
The work of navigation and translation is relatively easy. Drafters structure the
legislation so as to be meaningful to the audience, use manageable sentences and a
familiar vocabulary, and explain unfamiliar terms. They use techniques that give the
audience a bird’s-eye view of the structure. They create method statements that set out
the steps users must follow to achieve a particular goal. They include notes and other
devices to draw users’ attention to definitions and related provisions.
The work of explaining and personalizing the legislation is more daunting. Cer-
tain conventions currently used by drafters, notably preambles and purpose state-
ments, can help an audience to relate the text to their own circumstances by supplying
contextual information. Recently developed techniques, such as the use of metaphor,
examples, or a question-and-answer format, are meant to further facilitate effective
self-application. Because the discipline of legislative drafting is new, drafters have
much to learn before achieving anything like a satisfactory set of techniques. But in
principle, the knowledge necessary to develop those techniques can be acquired.
1. Access
The problem of access to the statute book is currently being addressed through
electronic publication.’ This approach makes the text of legislation available to every-
one who has access to the Internet. More important, because the technology used in
publishing legislation by electronic means resembles the technology used for other
purposes, the skills developed by citizens in searching for good vacation spots or
downloading music can be applied to locating and successfully navigating the statute
book.
2. Understanding
Once users have located the statute book, they must be able to recognize the pro-
visions that are relevant to them and the significance of those provisions. As tradition-
ally drafted, a statute tells a particular kind of story within the conventions of a genre;
like other literary texts it is laid out from left to right, from start to finish. Unlike other
literary texts, however, legislation is not taught in school and readers are unfamiliar
with the genre and its conventions. Most readers are unwilling to read it from start to
finish, or even from left to right. They want to go directly to the bits that concern
them, find the answers to their questions, and ignore the rest.
To respond to this reality, drafters try to lead users where they want to go while at
the same time ensuring that they notice everything that is relevant to their circum-
stances and have enough context to appreciate the significance of what they notice.
3 See T. Scassa, “The Best Things in Law Are Free? Towards Quality Free Public Access to Primary
Legal Materials in Canada” (2000) 23 Dal. L.J. 301.
2001]
R. SULLIVAN – THE PROMISE OF PLAIN LANGUAGE DRAFTING
103
For this purpose, drafters rely on tables and indexes, overviews and flow charts, ex-
planatory notes, discursive cross-references,’ and the like. Statutes that are published
online should be able to take advantage of the multimedia and interactive dimensions
of electronic communication to develop new aids. Through hyperlinking, for example,
an audience can be given immediate access to provisions scattered through the statute
book and to materials incorporated by reference. It can also be given access to re-
sources like legislative history, legislative evolution, administrative guidelines, case
holdings, or even academic comment.
3. Personalization
Most citizens read legislation not to discover the law for its own sake but to try to
make the law work for them. This realization has important implications for both
drafting and interpretation. Jorge Gracia writes that the function of all interpretation is
to produce appropriate acts of understanding in relation to a text, but what is appro-
priate varies depending on the interpretation’s purpose.! In “textual interpretation”, as
he calls it, the purpose is to understand the meaning intended by an author or a
meaning that is warranted by the text. Understanding, or identifying a “correct’
meaning, is an end in itself. In “non-textual interpretation”, the purpose is to under-
stand the implications a text has for something outside the text, that is, the import of
the text for a matter of interest to the interpreter.’
When the official interpreters of legislation apply a statute to facts, they normally
understand themselves to be engaged in what Gracia would call textual interpretation.
Their goal is to understand the law by ascertaining “the meaning” of the text, in effect
to “de-text” the law. When non-official interpreters apply legislation to facts, however,
they engage in what Gracia calls non-textual interpretation. Their goal is not to under-
stand the law for its own sake, but to personalize it, even colonize it, by ascertaining
how it optimally relates to their own interests, needs, circumstances, or preferences.
To assist citizens in this form of interpretation, drafters rely on concrete language
(in addition to, if not in lieu of abstractions), often in the form of examples and illus-
trations. In a paper-based format, the examples and illustrations must be set out
somewhere in the statute-within a section or subsection, in a footnote, or possibly in
a schedule or appendix. In electronic format, examples and illustrations can be offered
in a variety of forms, including audio and video links. In either case, the drafter at-
tempts to model the reasoning used by lawyers to personalize legal rules.
A discursive cross-reference tells readers what is in the text being referred to so they can judge
whether it is worth their while to pursue the reference.
SJ.J.E. Gracia, A Theory of Textuality: The Logic and Epistemology (Albany: State University of
New York Press, 1995) at 159-63.
6/bid at 164-65.
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 47
B. Audience Assessment
Traditionally drafters have come to know their audience by reading reported
judgments. However, since plain language drafters want to communicate with parlia-
mentarians and members of the public as well as judges, they make an effort to ex-
pand these limited horizons. A plain language drafting project typically begins with
an effort to systematically identify the anticipated users of legislation, followed by
consultation with representatives of those users. Later in the process, successive ver-
sions of the draft are circulated among the representatives for comment or they may
form the basis for a program of focus testing. The feedback received from these ini-
tiatives is analyzed and relied on in the preparation of each new draft.
For example, when the United Kingdom undertook a plain language rewrite of its
fiscal legislation, it began with a series of consultations designed to include “all rele-
vant private sector interests and, in particular, … those who represent users of tax leg-
islation.” ‘7 The rationale for this approach was explained as follows:
[S]ince the whole focus of the rewrite … is to make tax legislation easier to use,
an exceptionally high degree of user involvement is essential, making the re-
write, in effect, a joint venture. It is therefore particularly important to have ar-
rangements, planned from the outset, to ensure that the user’s perspective is
fully reflected at every stage of the project.’
In Canada, Human Resources Development Canada (“HRDC”) and the Depart-
ment of Justice (“DOJ”) have been working on a plain language rewrite of the Em-
ployment Insurance Act (“EIA”) since 1997. Early in the process a Plain Language
Advisory Group was established, consisting of representatives of business organiza-
tions, labour organizations, and advocacy groups. This group has been meeting two or
three times a year to review every aspect of the rewrite process, from restructuring the
EIA and devising a new format to changing the terminology.
At the formal meetings, the consultants are asked questions about specific provi-
sions or features of the text-what they think of drafting headings in the form of
questions, for example, or whether a particular flow chart is helpful and accurate.
Between meetings they are sent copies of work in progress and invited to comment on
anything, from their general impressions of the draft to the clarity of particular words
7 D. Salter, ‘Towards a Parliamentary Procedure for the Tax Law Rewrite” (1998) 19 Statute L. Rev.
65 at 68.
‘Inland Revenue, “The Tax Law Rewrite: The Way Forward” at c. 9, online: Inland Revenue
in Salter, ibid at 67.
‘ S.C. 1996, c. 23. Unless otherwise indicated, information in this paper about the EM Plain Lan-
guage Project is based on my own participation in it as part of research funded by the Social Sciences
and Humanities Research Council of Canada.
2001]
R. SULLIVAN – THE PROMISE OF PLAIN LANGUAGE DRAFTING
or sentences. The information gathered from this feedback is formally recorded, re-
viewed, and factored into subsequent drafts.
C. Reliance on Experts from Other Disciplines
Effective communication requires an understanding of how communication
works and what tends to hinder it or help it along. For this purpose, knowing the law
and the conventions of legislative drafting is not enough. Drafters need help from
other disciplines, particularly psycholinguistics and document design. They need to
see their profession established as a scholarly discipline. This entails investigating the
premises of their practice, developing knowledge about the type of communication
that is involved in using legislative texts, and proposing theories and carrying out re-
search.
As a first step in achieving these goals, plain language projects sometimes adopt
an interdisciplinary team approach in which specialists from other disciplines become
part of the drafting team. The drafter still does the writing, but every aspect of the text
reflects the ideas of the team.'” More frequently, consultants are hired to address par-
ticular problems or issues.
For example, when it came time to develop a format for the plain language ver-
sion of the EA, the HRDC/DOJ project hired a firm of document designers who not
only came up with an innovative format for the legislation but also gave presentations
and prepared reports explaining the reasoning and research on which its recommen-
dations were based. More recently, the English Legislative Language Committee of
the Legislative Services Branch, DOJ, commissioned reports from two scholars in
communications, one in English and the other in French, to assess the use of para-
graphing in the English and French versions of federal legislation.”
D. User Testing
One of the most striking features of plain language drafting is the current interest
in testing audience response. Historically, nearly all advice about effective writing was
based on the intuitions of the advice-giver. Currently, drafters want evidence that the
” This approach was used with success in Australia’s rewrite of its corporations legislation. See V.
Robinson, “Rewriting Legislation: Federal Australian Experiences” (Paper delivered to the Depart-
ment of Justice Continuing Legal Education Scholar Series, “Making Laws Easier to Read: What’s
the World Coming To?” 8 March 2001) at 4-5 [unpublished].
” See M.P. Jordon, “Paragraphing in Legislative Writing: Linguistic and Pragmatic Foundations”
(Research Report for the English Legislative Language Committee, Department of Justice, Canada,
March 2001) [unpublished]; C. Beaudet, “La lisibilit6 des 6num6rations verticales” (Research Report
for the Minister of Justice of Canada, December 2000) [unpublished].
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 47
techniques they are using actually do facilitate communication. For this reason, user
testing is an important element of many plain language projects.
Probably the best known example of user testing was developed by Phil Knight
and Joe Kimble in the course of their plain language rewrite of South Africa’s human
rights legislation.'” The Knight-Kimble team developed two simulations, which were
used to assess the ability of an audience to find, read, interpret, and apply the legisla-
tion. The first simulation was developed for professionals, who in this test were repre-
sented by 139 law students from four universities. The second was developed for a
group of 96 “lay people”. Members of this latter group were diverse in terms of gen-
der, age, first language, education, and race.
Knight reported the following results. First, “[c]ompared to the original, the re-
vised Act resulted in improved scores for every task involved in use of the text. The
improvement was seen in the use of the Act by legal professionals and lay readers
alike.”” Second, in achieving this improvement, getting rid of legalese and improving
sentence structure were important factors. However, the impossibility of eliminating
essential legal concepts from the text was a major impediment. As Knight observes,
“Unfamiliar legal concepts torpedo comprehension, and sink it without a trace.” 4
Third, although both professional and non-professional users did better with the re-
vised version, neither group did very well. Among professionals, only 50 per cent
managed to score 80 per cent or better on the test. Among non-professionals, only 30
per cent scored 60 per cent or better.'” As Knight concluded, “[T]he revised text made
great gains over the original … But clearly, there is a need for more work, more in-
volvement with readers, and more study of their needs.”‘ 6
As part of the plain language rewrite of the EIA, a consultant was hired to test
user response to two plain language versions.” In one version, the legal subject was
‘2 The testing is reported in P. Knight, Clearly Better Drafting: A Report to Plain English Campaign
on Testing Two Versions of the South Africa Human Rights Commission Act, 1995 (Stockport, U.K.:
Plain English Campaign, 1996).
3 Ibid. at 39.
‘4Ibid. at 38.
‘5 Ibid. at 40.
16 Ibid.
” This program of testing was funded in part by a grant from the Social Sciences and Humanities
Research Council of Canada, and was carried out by Vicki Schmolka and GLPi. A full description of
the testing and analysis of results is set out in GLPi & V. Schmolka, A Report on Results of Usability
Testing Research on Plain Language Draft Sections of the Employment Insurance Act: A Report to
Department of Justice Canada and Human Resources Development Canada (August 2000) [unpub-
lished]. Schmolka also conducted user testing on regulations made under the Explosives Act (R.S.C.
1985, c. E-17) dealing with the retail sale of fireworks. See V. Schmolka, Consumer Fireworks Regu-
lations: Usability Testing, TR1995-2e (Department of Justice Canada, 1995) [unpublished].
2001]
R. SULLIVAN – THE PROMISE OF PLAIN LANGUAGE DRAFTING
107
addressed directly as “you/vous”; in the other the legal subject was referred to in the
third person as “the claimant/le demandeur’. Testing was done with groups in four
categories: (1) General Public (English); (2) General Public (French); (3) Informed
Users (English); (4) Informed Users (French). Each subject completed a self-
administered questionnaire which tested ability to find and understand information in
the legislative text. After completing the questionnaire, subjects participated in a focus
group discussion.
Like Knight’s testing program, the EM test was designed to compare the experi-
ence of people using a plain language version to the experience of those using the cur-
rent version. In all, 146 people were tested, 101 in the General Public category and 45
in the Informed Users category. Subjects in the General Public groups were employed
persons who thought they might need to claim insurance benefits at some time in their
lives, but they had no recent experience with the EM nor had they read any type of
statute or legal text within the past year. They included a mix of genders, ages, occu-
pations, and levels of education and income. Subjects in the Informed User groups in-
cluded Canada Employment Centre staff, human resource and benefit specialists, and
professionals such as lawyers, employment advocacy groups, community assistant
groups, and the like. All had experience advising people about employment issues and
were familiar with the current E/A.’9
Overall, the results were similar to those reported by Knight. Groups working
with both plain language versions performed better than groups working with the cur-
rent version. The former found information more quickly and were able to answer
questions about it more accurately. They also reported less distaste and more confi-
dence in their ability to use the legislation. However, they still found it frustrating to
try to find things in the EM and the text remained difficult to understand. 9 Their per-
formance in the questionnaire part of the testing reflects this subjective impression.
Neither the general public nor informed users did particularly well on any version of
the legislation.?0
On the question of direct address versus third-person reference, the results were
largely inconclusive. While many people thought using “you” or “vous” was helpful
and user-friendly, a significant number reacted strongly to it, finding it aggressive or
patronizing. Others had trouble figuring out who “you” or “vous'” was supposed to be.
Moreover, those who worked with the third-person version performed slightly better
in finding and understanding the answers to questions. For these reasons, the testers
recommended the third-person approach.’
‘9GLPi & Schmolka, ibid. at 2-3.
I9Ibid at 5-10.
20This is my own assessment based on reading GLPi & Schmolka, ibid.
21 Ibid at 8-9.
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 47
Given the small numbers involved, the results of this type of testing are not statis-
tically significant. They do not permit analysts to explore possible correlations be-
tween features like gender or age and responses to the legislation. Generally, they do
not attempt to assess the impact of particular features. The results do suggest, how-
ever, that plain language drafting is an improvement over traditional legislation, but if
direct communication is the goal, plain language drafting has a long way to go.
II. Writing for Divergent Audiences
Although there are exceptions, legislation is generally addressed to everyone. The
typical legislative sentence asserts a rule in the following form: “Any person who
[comes within described circumstances or meets described conditions] is entitled
to/obliged to/liable to [something].” Although the circumstances and conditions set
out in the restrictive clause limit the scope of the rule, the subject of the rule is “any
person”, a description that encompasses all of us.
Even though legislation is addressed to everyone, historically it has not been
written for everyone. It has been written for the legal profession with the primary fo-
cus on a somewhat idealized conception of superior court judges. This audience has
above average intelligence and sophisticated reading skills; its members are highly
educated and unusually well-informed. They have an extensive knowledge of law and
legal values. Furthermore, they understand the conventions used by drafters in pre-
paring legislation and they have mastered the rules and techniques of statutory inter-
pretation. Most Canadians, including some lawyers and judges, would have trouble
locating themselves in this group portrait.
But what are the alternatives? If drafters do not write for judges, for whom should
they write? In this part, I canvass four possibilities: writing for everyone, writing for
the audience targeted by Parliament, writing for the least experienced, and writing for
actual readers. I express skepticism about the possibility of directly communicating
the law to the general public and I consider whether it makes sense to write for an
audience most of whose members will never read what is written. I end by urging
drafters to write for the audience that has been singled out by the legislature, or in the
absence of such an audience, for the most vulnerable groups affected by the legisla-
tion.
A. Writing for Everyone
Many people assume that drafting in plain language means writing for every-
one-in fact, writing for that broad, undifferentiated audience to which traditional
legislation is addressed. Drafters who adopt this approach sometimes attempt to un-
derstand this audience in terms of “average” ability, “ordinary” knowledge, “typical”
background, and the like. This is the approach of Dennis Murphy, an early advocate
of plain language drafting. He writes:
2001]
R. SULLIVAN – THE PROMISE OF PLAIN LANGUAGE DRAFTING
109
There has been much discussion over the years as to the differing needs of Par-
liament, lawyers, the judiciary, professional groups, special groups, or the pub-
lic itself.
I submit that the correct principle is that, while all these stakeholders
should be catered for according to the nature of the legislation, the overriding
concern should be to ensure that the person in the street, the ordinary person,
should be thought of as the ultimate consumer of legislation. Why shouldn’t the
ordinary person have full and real access to legislation? …
Legislation should be written so that it is feasible for the ordinary person of
ordinary intelligence and ordinary education to have a reasonable expectation
of understanding and comprehending legislation and of getting the answers to
the questions he or she has.”
In my view, conceptualizing the general public in terms of what is ordinary or av-
erage or reasonable is not an effective strategy. The “ordinary person” to whom Mur-
phy refers does not in fact exist. The drafter is therefore obliged to make this person
up on the basis of intuition and imagination. Although the average person made up by
plain language drafters is not a lawyer, he is probably male, of European descent,
middle-aged and mainstream in his thinking, and comfortably ensconced in the mid-
dle class. The differences between this person and the judge for whom traditional
legislation is written are minimal. This so-called average person lacks legal knowl-
edge and abhors legal jargon, but he could easily golf with the judge.
An alternative way of thinking about “everyone” is to survey the mass of people
to which the term refers and classify them based on characteristics that are thought to
be relevant. The virtue of this approach is that it acknowledges that real people differ
from one another in ways that affect communication. Duncan Berry, who has done
considerable research on audience identification, illustrates this approach. He writes:
Audience is a broad concept. In order to understand audiences, legislative
counsel must first ascertain who will read their legislation and how they will
use it. … [R]eaders may vary widely, so it is necessary to identify the specific
characteristics of the various audiences in order to make the legislation accessi-
ble to readers. Similarly, legislative [counsel] must identify how their audiences
will use the legislation.
Legislative counsel can identify [their] audiences if they think of all who
will potentially read the legislation or whose activities it will control. When
those audiences have been identified, the legislation’s relationship to its readers
should be considered. An audience may be friendly, that is, share interests
2 D. Murphy, “Plain English-Principles and Practice” (Paper delivered to the Conference on Leg-
islative Drafting, Canberra, Australia, 15 July 1992) at 4 [unpublished].
110
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 47
similar to the policy formulators, or hostile, that is, have interests that conflict
or potentially conflict with those of the policy formulators.
Another factor to consider in identifying the audience is its education and
experience.”
Berry suggests that the audience for most legislation would include judges and law-
yers, the bureaucrats and officials who administer and implement the legislation, the
persons affected by it, and the advisors of those affected. After canvassing several
methods that can be used to analyze these audiences, he concludes:
Audience analysis should include a comparison of the author and the audience
and an assessment of their respective knowledge, values and beliefs about the
subject matter. A comparative analysis can put legislative counsel in a more in-
formed position to make visual and verbal decisions that may bridge the gap
between themselves and their audience.24
Clearly Berry appreciates the difficulty of trying to write for real persons as opposed
to “the average person”. He repeatedly draws attention to the different assumptions,
abilities, interests, and needs that different categories of audience bring to legisla-
tion-along with different reading strategies. He acknowledges that these differences
operate as barriers to communication. Despite all this, he remains optimistic. At the
end of the day he believes that drafters can get to know these different audiences well
enough to communicate with them directly and effectively. He believes that the needs
of these audiences can all be accommodated within a single document-not fully ac-
commodated perhaps, but accommodated to an extent that justifies trying to write for
them all simultaneously.
I myself do not think it is possible for drafters to bridge so many different gaps
simultaneously. It may be possible to draft legislation that is easier for many different
audiences to use-professionals and non-professionals, the highly literate and not so
literate. User testing in South Africa, Hong Kong, New Zealand, and Canada supports
this claim. However, it is important to notice the difference between drafting legisla-
tion that is easier for different audiences to use and writing for different audiences. If
audiences in fact have different needs and interests, or bring a different knowledge
base to the legislation, drafters must either shift back and forth among the several
audiences, accommodating sometimes one group and sometimes another, or they
must single out a primary audience whose needs become their primary (although not
exclusive) concern. For reasons explained below, I think the latter approach is better.
23 D. Berry, “Audience Analysis in the Legislative Drafting Process” (June 2000) The Loophole: J.
Commonwealth Ass’n Legis. Couns. 61 at 62.
24 Ibid. at 67.
2001]
R. SULLIVAN – THE PROMISE OF PLAIN LANGUAGE DRAFTING
B. Writing for the Audience Targeted by Parliament
Having to draft legislation for audiences with competing interests, divergent
backgrounds, and unequal power is a challenge that drafters face on a daily basis. The
problem arises in virtually every statute that affects the general public, yet it is rarely
noticed or discussed, even in the context of plain language rewrites. For example, Mi-
chael Jordan, commenting on a bill to replace Ontario’s Residential Rent Regulation
Act, 1986,’ writes that the bill is “of great importance to the general public in On-
tario-and specifically those who rent, rather than own, property “‘2 He concludes that
the new legislation “needs to be written in a form that renters can understand and act
on, as they are directly affected by the provisions of the Act.”‘
As Irene Gendron points out, a rent regulation act sets up a scheme of rights and
obligations for both renters and landlords. Why does Jordan assume that renters are
the primary audience? Here is Gendron’s speculation:
Perhaps Jordan is making a judgment that renters are the “lowest [common]
denominator” in the audience and that writing for their level of understanding
will automatically allow landlords to understand the statute. I suspect, however,
that it is more of a political decision-renters are viewed as being in a less fa-
vourable position than landlords and therefore are the “underdogs” … The
situation is equalized by writing the legislation for them rather than for the
landlords.’
Gendron rightly characterizes the decision to make renters the primary audience as a
political one. She goes on to point out how such a decision can affect the way the
legislation is drafted:
[S]pecifically directing the legislation to the renter is different than specifically
directing it to landlords. For example, if we … give prominence to the informa-
tion that is of most importance to the renter (as the preferred audience), then
those items would be placed first and items that are of more significance to
landlords or bureaucrats would be placed in positions of lesser prominence.
Also, headings and marginal notes would be written for the renter as reader.
For example, “How much of a rent increase will I face each year?” If the Act
were aimed at landlords, the heading might be: “How much can I raise the rent
each year?” If kept neutral, perhaps so as to engage both groups of readers (as
2 R.S.O. 1990, c. R.29, as rep. by Rent ControlAct, 1992, S.O. 1992, c. 11, s. 137(1).
” M.P. Jordan, “Plainer Legal Language: Definitions and Requirements in Acts” (1994) 24 J. Tech-
nical Writing & Comm. 333 at 336.
2Ibid.
I .V Gendron, “Can a Statute Be All Things to All People? The Reality of Audience-Based Leg-
islative Drafting” (LL.M. Major Research Paper, University of Ottawa, 31 August 2000) at 36 [un-
published].
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 47
well as tribunals and bureaucrats), the heading might read “Amount of allow-
able yearly increases” or “How much rent increase is allowed?” ’29
In commenting on the Jordan article, Gendron wonders whether the decision to
single out and write for a particular audience can properly be made by the drafter,
given its political character. In my view, drafters do not really choose the primary
audience, but rather identify the primary audience chosen by Parliament. They do this
by interpreting their instructions, a task that is central to a drafter’s day-to-day job.
Either they interpret the cabinet memorandum and the more detailed instructions sup-
plied by department or agency officials or, in the case of a rewrite, they interpret the
existing legislation. In either case they canvass the interests affected by the legislation
and establish how these interests are balanced or ranked in the intended scheme in or-
der to appreciate the political choices that Parliament will be making. Their job is to
find ways to embody and communicate these political choices in the text, and one
way of doing this is to signal the primary audience for which the legislation is written.
This is the approach adopted by Gendron herself, along with her co-drafter Philippe
Hall6e, in their initial plain language draft of the EM. Here is her description of their
early deliberations:
We started by listing the current users of the legislation. Empirical research
was not available but it was not difficult to identify the current users of the Act,
at least in a general way:
o
Tribunals
o Courts
o
o
o
Lawyers
Employers, employees (claimants) and their lay and professional
representatives and advocates
staff of HRDC and Canada Customs and Revenue Agency
We spent a great deal of time during the initial phase of the project analyz-
ing the statute with a view to re-structuring it. We asked ourselves questions
such as-which information should be put first? Which is the most important
information in the Act? Which segments of information belong together? As
we struggled with these issues, it became clear to us that the answers were dif-
ferent depending on the point of view we took … that is, which … audience we
adopted. Although we did not analyze it in any depth at the time, it seemed to
us that the common denominator throughout most of the Act was the “claim-
ant”-a person who wants to claim, or is claiming, benefits under the legisla-
tion. The thrust of the Act is, after all, to provide a scheme of income replace-
” Ibid. at 36-37.
2001]
R. SULLIVAN – THE PROMISE OF PLAIN LANGUAGE DRAFTING
ment for the claimant-the Act exists for that purpose … although a portion of
the Act is specifically aimed at employers.”
After consulting with other members of the team, Gendron and Hall6e decided that
the plain language version of the EA should be written for claimants-that is, organ-
ized and presented from their perspective.
This decision did not mean that every provision in the EM would have to be
written for claimants. Much of the EA deals with the duties of employers or with ad-
ministration of the scheme, and in drafting these parts, the drafter would try to com-
municate directly with employers or with administrators. It is only when conflict
arises and the drafter must choose between competing interests or needs that prioriti-
zation becomes necessary. My argument is that when conflict does arise the drafter
should favour the primary audience targeted by Parliament.
C. Writing for the Least Experienced
The problem of choosing among different audiences was handled in a somewhat
different way by the drafters who did the plain language rewrite of South Africa’s
Human Rights Commission Act, 1994.” As Knight reports, they envisaged the audi-
ence for this legislation as including parliamentarians, judges and lawyers, commis-
sion members and staff, police and other enforcers, advocates, community service
providers, business people, teachers, labour leaders, ministers, and ordinary citizens. 2
Like Berry, they believed they could draft a text that would improve readability and
usability for all these groups; in fact, this was an important goal of the project. But
they acknowledged the impossibility of writingfor all these groups simultaneously.
The problem of competing needs and interests arose at the outset, when the draft-
ers sat down to devise a new structure for the legislation that would reflect the logic
and interests of the reader. This is an early and fundamental step in plain language
drafting. Knight writes:
Bearing in mind that there are different groups of readers, with different inter-
ests and purposes in reading the document … [i]t is inevitable that they will not
all share the same logical approach to arranging the subject matter.
We resolve those competing interests in favour of the least experienced
readers, and their primary interests, providing a logical transition [from one
“Ibid at 52-53.
“‘ No. 54 of 1994, online: Parliament of South Africa
” Knight, supra note 12 at 12.
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 47
subject to the next] that [was meant to] reflect the hierarchy of interests of the
most common readers, or at least the most frequent unofficial readers.33
Knight’s solution to the problem of competing audiences, in this context at least, was
to write for the least experienced readers. He does not explain why he chose to write
for this audience, but the choice is intriguing.
Perhaps the deciding factor was the size of the group. Perhaps Knight meant to
suggest that drafters should write for the audience with the largest number of readers.
In the passage quoted above, he identifies the least experienced readers with “the most
common readers” and “the most frequent unofficial readers”. It is more likely, how-
ever, that the key factor was not the size of the audience but its members’ familiarity
with legislation or their level of literacy. Probably he meant to suggest that drafters
should write for the readers who need the most help in dealing with a legislative text.
If the text can be read and understood by the least experienced readers, then presuma-
bly it can be read and understood by more sophisticated readers as well. However, it is
also possible that Knight and his team were thinking about power. The least experi-
enced readers are probably the most vulnerable members of society and, coinciden-
tally, the members least likely to have access to professional help. Perhaps Knight
meant to suggest that drafters should write for the audience with the least power in the
system. After all, this group most urgently needs the services offered by the plain lan-
guage drafter.
In my view, writing for the most vulnerable groups that will be affected by par-
ticular legislation is the appropriate default position for drafters. I would suggest that,
in the absence of instructions to the contrary, drafters are not only entitled to write for
this audience but may even have a professional obligation to do so. As mentioned at
the outset, a drafter’s primary job is to transform the instructions received from Cabi-
net into legally effective legislation. An important part of this job is making sure that
proposed legislation complies with formal constitutional constraints and with estab-
lished constitutional values. Given the historical importance placed on equality and
equal access to the benefit of the law, drafters are well within their mandate in at-
tempting to minimize the power differences between those who have the benefit of
professional intermediaries and those who do not.
D. Writing for Actual Readers
In the Canadian and South African projects described above, when drafters were
forced to single out their primary audience they chose to write for a particular sub-
category of the persons whose lives would be affected by the legislation. In other
plain language projects, however, drafters have chosen to write for the audience that is
” Ibid. at 9.
20011
R. SULLIVAN – THE PROMISE OF PLAIN LANGUAGE DRAFTING
115
most likely to read the legislation. It is easy to see their point. What good does it do to
write for people who will never read what you have written? Unless there is an an-
swer to that question, writing for an audience of renters, or claimants, or the most
vulnerable groups in society is a waste of time.
Writing for the actual readers of legislation is the approach taken in Australia,
which has been on the cutting edge of plain language drafting since the mid-1980s. To
date it has completed eight major plain language rewrites, beginning with social secu-
rity legislation in 1989. Other rewrites have dealt with income and sales tax, offshore
mining, care of the aged, export market development, corporate governance, and the
public service.’ Vimce Robinson, who has been involved in several of these projects,
explains the Australian approach to audience:
Our rewrites have been pragmatic rather than ideological. We have aimed,
by and large, to make the legislation more readable and easier to use for its ac-
tual current users. We have not tried to make the legislation accessible to people
who, from a practical point of view, we believed were highly unlikely ever to
read the legislation….
The actual current user group is usually a challenging enough group to ca-
ter for. Often it will include not only high-powered legal advisers or account-
ants but also professionals with lower level skills, the government officials who
administer the legislation and community advice bureau staff who give people
advice on their rights.
To some extent, you can, in aiming for an unrealistic audience, do a dis-
service to the actual users. For example, if your actual users are professional
advisers who already have some familiarity with the area and you put in a lot of
aids for the benefit of uninformed lay users, you may alienate or inconvenience
the actual users.
We have not aimed to make the legislation readable by the average citizen.
It is unusual for the average citizen to encounter a scheme through the legisla-
tion itself. The average citizen is more likely to encounter the scheme through
brochures, notices, forms or websites.
That said, the average citizen does get the benefit of the improvements to
the legislation that are introduced for the benefit of the actual users. The things
that make life easier for the actual users … also tend to make life easier for
other potential users.”
34See e.g. the series of fiscal acts introduced in 1999 beginning with A New Tax System (Aged Care
Compensation Measures Legislation Amendment) Act 1999 (Austl.), No. 58, 1999 and ending with A
New Tax System (Tax Administration) Act 1999 (Austl.), No. 179, 1999.
3 Robinson, supra note 10 at 8-9.
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 47
Robinson is certainly right when he suggests that writing for an audience of ac-
tual users, given its diversity, is challenge enough. He is also right to suggest that the
needs of expert and non-expert readers often come into conflict, and that expert read-
ers are probably better served by a text prepared to meet their needs. The differences
between these categories of audience have been documented in other contexts. Karen
Schriver writes:
As audiences become more specialized and more educated in technical ar-
eas, they expect texts that are not designed “for everyone:’ but rather, are tar-
geted to their particular needs. In many industrial and corporate document de-
sign contexts (such as the computer, electronics, and appliance industries),
writers must tailor their texts to very particular audiences. The ability to adapt
texts for audiences such as novices, intermediates, or experts is rapidly be-
coming a requisite skill in industry.36
Finally, Robinson is probably right to suggest that non-expert readers prefer to find
out about the law through intermediaries rather than by reading legislation. As
Schriver also writes, “What we know now is that most people choose to read and to
keep reading only when they believe there will be some benefit in doing so and only
when they cannot get the same information in easier ways (for example, by asking
someone else).”3 It is obvious that no one (not even a lawyer) reads statutes for pleas-
ure. If members of the public who are affected by legislation can afford to hire a pro-
fessional, they will do so. If they cannot afford their own professional, they will rely
on the services provided by community workers or government bureaucrats in the
form of brochures and pamphlets, advertisements, answers to inquiries, and the like.
Despite its many insights, there are problems with Robinson’s analysis. For one
thing, he is too quick to dismiss the possibility that some users of legislation may
want to bypass the intermediaries available to them and read the legislation for them-
selves. At a meeting of the Employment Insurance Act Advisory Group, held on 8
March 2001, representatives of labour groups described two situations in which
workers currently do read the EIA unassisted by intermediaries. The first was de-
scribed by a representative of the British Columbia and Yukon Territory Building and
Construction Trades Council, who explained that in the construction industry, and
other industries like it, unemployment is recurrent and rooted in the circumstances of
the local economy. In most localities, a particular worker becomes an authority on
employment insurance law. That person has up-to-date copies of the EIA and the
regulations, is familiar with their content, interprets the legislation, and dispenses ad-
‘6K.A. Schriver, “Plain Language through Protocol-Aided Revision” in E.R. Steinberg, ed., Plain
Language: Principles and Practice (Detroit: Wayne State University Press, 1991) 148 at 151-52.
” K.A. Schriver, Dynamics in Document Design: Creating Text for Readers (New York: John Wiley
& Sons, 1997) at 166 [hereinafter Schriver, Dynamics].
2001]
R. SULLIVAN- THE PROMISE OF PLAIN LANGUAGE DRAFTING
117
vice to his or her peers, thereby becoming the main source of information in the par-
ticular subcommunity on the rights and obligations of workers under the E!A.
Bernard Jackson would call this kind of subcommunity a semiotic group, that is, a
group of people who share a context that differs from the contexts of others in ways
that may lead them to infer different meanings from the same legal text. Jackson
writes: “[A] semiotic group is a group which makes sense (here, of law) in ways suf-
ficiently distinct from other such groups as to make its meanings less than transparent
to members of other groups without training or initiation “‘3″ To communicate directly
and effectively with a given semiotic group, the drafter would have to receive training
or initiation into the group; this is what consultations and user testing are all about.
But of course it is impossible for a drafter to master the context of every semiotic
group that will be affected by legislation like the E!A.
The peer interpreters described above offer an attractive solution to this problem.
They are ideally placed to mediate between the government and a small subcommu-
nity, and in particular to personalize the legislation for members of their group. The
peer interpreter saves the group from being at the mercy of the bureaucrats who pre-
pare government brochures and answer inquiries from the public. Unlike lawyers and
accountants, government bureaucrats owe no duty of care or competence to those they
serve; their first loyalty is to the government. This makes them untrustworthy from the
point of view of most subjects. Even if government bureaucrats wished to act as ef-
fective intermediaries, they would not find it easy to personalize the legislation-to
adapt it to the local group in a way that serves the latter’s interests. Because peer in-
terpreters understand local circumstances and share the interests and perspectives of
their group, their efforts to adapt the legislation to their own situation have the effect
of personalizing it for the entire group.
Other situations in which workers choose to avoid the services of intermediaries
were described by a representative of the Canadian Labour Congress. He referred to
workers whose claim for benefits has been denied or to those who find themselves in
trouble because of an alleged fraud. Most workers in this position cannot afford a
lawyer,”3 and they have every reason to distrust the information that issues from the
government’s intermediaries. In these circumstances, workers who want to determine
” B. Jackson, “Legislation in the Semiotics of Law” in H. van Schooten, ed., Semiotics and Legis-
lation: Jurisprudential, Institutional and Sociological Perspectives (Liverpool, U.K.: Deborah Char-
les, 1999) 5 at 6.
” This turns out to be a blessing in disguise. According to a representative of the Canada Employ-
ment Commission, the lawyers who appear before the commission on behalf of workers usually do
little to assist their client’s cause. They are unhelpful because they are unfamiliar with the legislation
and they do not understand how the system works. It appears that the costs of acquiring the knowl-
edge necessary to do a good job of personalization exceed any benefit that could be derived from this
client base.
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 47
whether they are being treated legally are driven to the EM itself. They may receive
some help from a union advisor or a community worker, but because the ratio of
helpers to persons needing help is extremely low, most of the personalization will
have to be done by the worker himself or herself.
These examples suggest that within the category of people whose lives are af-
fected by legislation, there will always be some who bypass intermediaries and go
straight to the act itself. Although the number of such readers is small, they should not
be discounted for they have the best claim to the services of the plain language drafter.
E. Conclusion
The problem with writing for the general public is that it is impossible to person-
alize the legislation for such a diverse audience. The communities and subcommuni-
ties that constitute the general public have different needs, interests, and expectations;
they comprise different semiotic groups. This makes it impossible to communicate di-
rectly and effectively with all of them at once. By trying to be all things to all people,
the drafter risks disappointing everyone.
The problem with writing for those who will actually read the legislation is that
this audience consists of professional advisors. Drafters thus wind up writing for the
very intermediaries whose services they are supposed to render unnecessary. Obvi-
ously there is nothing wrong with wanting to improve communication with those who
must read legislation as part of their job. If the drafter is successful, professionals will
be able to do their work more efficiently. An improved text has the potential to reduce
uncertainty and improve compliance, and it is almost certain to save someone time
and money. But it is not designed to help outsiders enter the walled fortress. In
choosing to write for the people who are already inside, the plain language drafters
are no longer providing free translation, navigation, and explanation services to those
who cannot afford their own professional. They are now serving the audiences who
need them the least.
The advantage of writing either for the audience targeted by Parliament or, in the
absence of a clear target audience, for the most vulnerable groups affected is that it as-
sists those who have the best claim to assistance. It also sends an important message
to the official interpreters of the legislation, a message that in my view has significant
transformational potential. I would argue that the message sent by writing legislation
for the targeted audience or the most vulnerable groups is sufficient to justify writing
for these audiences even if most of their members never actually read the legislation.
This potential is explored in the following part.
Ill. The Impact of Plain Language Drafting
One of the major concerns in any plain language project is how the official inter-
preters of the legislation will respond to change. What will they make of the new
2001]
R. SULLIVAN – THE PROMISE OF PLAIN LANGUAGE DRAFTING
119
vocabulary, the shorter sentences, and novel features such as flowcharts, explanatory
notes, and examples? Those who oppose plain language drafting emphasize the costly
uncertainty that may result from changing the words of the text. They fear that matters
long ago settled will have to be litigated again. They also fear that interpreters will see
change where none was intended. An interpreter who is used to seeing “A person
shall” in legislation may be troubled by the words “A person must” and may conclude
that the legislature would not have changed such well-established wording if it did not
intend to change the meaning.’ In my view, this preoccupation with the meaning of
particular words and expressions misses the subversive potential of plain language
drafting.
I believe that an obvious attempt to communicate with non-professionals sends a
novel message to official interpreters about whom the law is for and the perspective
from which it should be interpreted. I also believe that several aspects of plain lan-
guage drafting have the potential to undermine the positivist assumptions on which
both drafters and interpreters rely to justify practices that serve the status quo. At the
least, plain language drafting “problematizes” the text and interpretation. It suggests
that there may be more to the business of interpretation than simply decoding legisla-
tive sentences. Potentially, it draws attention to the way law, including statute law, is
made up on an ongoing basis through its personalization in the daily life of citizens.
Before exploring these potentials, I should explain why I think it is important to
focus on official responses to plain language legislation. After all, the whole point of
plain language drafting is to communicate directly with the subjects of legislation and
to focus on their interests and needs. Why should we concern ourselves with judges
and police officers and bureaucrats?
The answer, of course, is that official interpreters have more power than the sub-
jects whose lives are affected by the law. If there is a dispute about how legislation
relates to particular facts, the dispute is resolved in accordance with the views of those
who have the most power in the hierarchy of interpreters. I may claim an entitlement
to an insurance benefit based on my own highly plausible interpretation of the EM,
but if the bureaucrat who processes my claim has a different interpretation, that inter-
pretation will prevail. Of course, the bureaucrat’s interpretation is subject to the
Commission, whose view is subject to the Umpire, and so on up to the Supreme Court
of Canada. In matters of interpretation, he who laughs last definitely laughs best.
Whatever one thinks of judicial review, given the reality of legal hierarchy in Western
democracies, the interpretations of the courts cannot be ignored.
‘a Although this illustration seems silly it is based on a case in which the judge held that “must” is
more imperative than “shall”. See Lovick v. Brough (1998), 36 R.F.L. (4th) 458 at para. 7 (B.C. S.C.),
online: QL (BC).
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 47
A. The Plain Language Message
As noted above, the courts are the primary audience for which legislation has
been written in the past, and current interpretive rules and practices are grounded in
this assumption. If you are a judge who is called on to interpret legislation that has
been written with your linguistic competence, your knowledge base, and your cultural
background in mind, interpretation is a warm bath. Everything you need to interpret
the legislation successfully is already inside your head. That is why statutory inter-
pretation does not require evidence. Courts take judicial notice of everything, from the
ordinary meaning of language to common sense, from the norms of justice and right
reason to the preferred policies of the legislature.
Plain language drafting does not directly attack or criticize this reliance by judges
on what they already know, but it is likely to get them thinking about it. By fore-
grounding issues of language and communication, the drafter invites interpreters to
become self-conscious about their work. This foregrounding occurs in a variety of
ways. First, and quite importantly I think, statutes drafted in plain language do not
look like ordinary legislation. The plain language version of the EIA, for example,
looks like a user’s guide to a software program or instructions for assembling a bicy-
cle. Judges are going to suspect right away that this legislation has not been written
for them, that an attempt is being made to communicate with a non-legal, non-
professional audience; in case judges might somehow miss the implicit message, the
first paragraph of the plain language version, under the title “Guide-How to use this
Act”, is explicit on the question of preferred audience:
The information in this Act is organised to help people who are claiming un-
employment insurance benefits. The most important information for claimants
is found at the beginning of the Act.”
In various ways a message is sent to interpreters: in this enactment the legislature is
attempting to communicate directly with its audience, and that audience is not com-
posed of judges and lawyers.
In my view, this message invites interpreters to reflect on their current practices
and to question the assumptions on which those practices are based. In particular, it
invites them to reconsider the assumption that ordinary meaning and common sense
are the same for everyone, that we all belong to the same semiotic group. By writing a
human rights act for the least experienced members of society or social welfare legis-
lation for potential claimants, a legislature directs interpreters to a particular context
and perspective, and suggests that interpretation should be grounded in that context
4, Working draft of the E/A Plain Language Project. The project is described above, text accompa-
nying note 9.
2001]
R. SULLIVAN – THE PROMISE OF PLAIN LANGUAGE DRAFTING
121
and perspective. It thus acknowledges the gap between official interpreters and those
audiences and makes it the responsibility of the interpreters to bridge the gap.
This directive has special implications for judges. It suggests that judges need to
develop a non-textual, audience-based approach to interpretation that complements
the plain language drafter’s efforts to make legislation accessible to non-professionals.
More precisely, it suggests that it is properly the business of judges to ensure that the
legislation is appropriately personalized for its primary audience.
The approach recommended here entails significant change. Instead of relying on
judicial notice of meaning and their own common sense, judges will have to receive
and assess evidence about the audience for which the legislation was written. They
will have to master an alien context and perspective. The evidence required for this
purpose could come from experts in sociolinguistics or related fields; it could also
come from the audience itself. A recent Ontario case, Stewart v. Canada Life Assur-
ance,2 illustrates the sort of thing I have in mind.
The issue in Stewart was whether a widow was entitled to be paid as beneficiary
of her husband’s life insurance policy. The insurance company alleged that the hus-
band had made a material misrepresentation in the insurance application and on this
ground refused to pay. The husband had suffered from ulcerative colitis since age
thirteen. Asked on the insurance application whether he had ever been diagnosed as
having “[d]iabetes, kidney or liver disease or any other disease of the stomach, intes-
tines, rectum, bladder, prostate, or reproductive organ,” he answered no. 3
Medically speaking, Mr. Stewart was wrong-colitis is a disease of the lower or
large intestine. From the point of view of persons suffering from colitis, however, this
was not so obvious. An Angus Reid poll, commissioned by the widow in support of
her claim, showed that most colitis sufferers think that they have a colon disease and
that the colon is distinct from, rather than part of, the intestines.
The trial judge was hostile to this evidence. In his view the poll merely showed
that “ordinary Canadians” have limited knowledge about the anatomy of the intestinal
tract and he did not need evidence to tell him that. “I am perfectly entitled to take ju-
dicial notice of it,’ he wrote.” In my view, there are two things wrong with this re-
sponse.
First, it shows how judicial notice permits judges to rely on their beliefs and im-
pressions without having to test or justify them. Crucial assumptions are taken as es-
tablished on the basis of personal conviction, which might be based on anything at all.
42 (1999), 14 C.C.L.I. (3d) 178, [2000] I.L.R. 1-3792 (Ont. Sup. Ct.) [hereinafter Stewart], aff’d
[2000] OJ. No. 2970 (C.A.), online: QL (OJRE).
43bid. at paras. 28, 30.
4TIbid. at para. 32.
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 47
Second, it confounds the response of colitis sufferers with the response of “ordinary
Canadians”. Although the poll was addressed to the general public and participants in
the poll were randomly chosen, the evidence tendered by counsel focused exclusively
on the response of people suffering from colitis. Colitis sufferers differ from “ordinary
Canadians” in having a greater interest in and more information about the disease.
That is what makes their response potentially more relevant. For purposes of under-
standing the question on the insurance application, colitis sufferers form a distinct
semiotic group.
In tendering evidence about the understanding of colitis sufferers, counsel for the
widow implied that the audience for which the question was drafted consisted of peo-
ple suffering from the diseases mentioned. Since this was the primary audience, in the
event of conflict its understanding should prevail over the understanding of other
groups-such as doctors or insurance companies or the public at large. In effect,
counsel was urging the court to take an audience-based approach.
My argument is that drafting legislation in plain language so as to single out a
particular audience is at the least a justification to take an audience-based approach
and, properly understood, is a directive to do so. In audience-based interpretation, the
measure of correctness is whether the interpreter has personalized the legislation in an
appropriate way.
B. The Death of Positivism?
Perhaps the most interesting dilemma arising out of plain language drafting is
how to deal with the “bells and whistles” that are a prominent feature of many plain
language statutes-the summaries and overviews, titles and headings, explanatory
notes, process statements, flow charts, examples, and illustrations. These features
clearly are part of the statute. But are they part of the law? Is the law co-terminus with
the text? And if so, what counts as text?
At one time, the only parts of a statute considered to be enacted law were the
words of the enacted provisions. Other parts-like headings and preambles-were in-
side the statute but external to the law. They were not text, but context. As such they
could be relied on to assist interpretation, but only if the text itself was ambiguous or
otherwise unclear. Finally, aids like marginal notes (and punctuation!) were neither text
nor context but mere editorial gloss. They had neither legal force nor interpretive value.
Over the years these distinctions have blurred and their basis has been largely for-
gotten. However, they have never quite disappeared. Courts are still inclined to treat
the numbered provisions of statutes as the text that declares the law and to give a
lesser status to headings, titles, and marginal notes. Chances are they will respond in
much the same way to the innovative techniques of plain language drafting. Is this re-
sponse appropriate? Should drafters tell interpreters to adopt a different response?
2001]
R. SULLIVAN- THE PROMISE OF PLAIN LANGUAGE DRAFTING
123
There are practical reasons why drafters might want to treat certain features of
plain language drafting as context rather than text and assign them an inferior status.
Consider the use of examples illustrating the application of one or more legislative
rules to a particular fact pattern. Examples of this sort are costly to prepare and costly
to monitor and maintain. Whoever prepares them must fully understand the legislation
and the context in which it will operate and must have sufficient imagination to see
the ramifications of each example. And preparation is only the beginning. So long as
the legislation remains in force, the examples must be monitored. Even a minor
amendment of a provision may require adjustments to several examples. Similarly,
changes in the operational context of the legislation may affect the import of an ex-
ample. Someone has to assess the impact of any changes and prepare appropriate re-
visions. In all this work, there is significant risk of oversight and mistake.
Other concerns arise as well, having to do with uncertainty and loss of control.
Because examples are vivid and emotionally engaging, they have the potential to hi-
jack debate. The fear is that legislators and interpreters will be distracted by the ex-
amples and ignore the rules. Furthermore, examples can be misused. Based on ejus-
dem generis reasoning, they may be used to inappropriately narrow the scope of a
benefit or prohibition that is set out in general terms in the numbered provision. Con-
versely, based on reasoning by analogy, they may be used to justify extending a bene-
fit or prohibition to other analogous fact patterns even though these fact patterns are
outside the language used in the numbered provision.
Underlying these concerns is a fundamental assumption about the nature of law.
In nearly all discussions of examples and similar “aids”, it is assumed that the law re-
sides in the words of the numbered provisions, which are drafted in the form of general
rules. These rules are drafted first, and once they reach their final form, the law is fixed:
it is fully and definitively captured in the language of “the text’. Everything the drafter
does afterwards is then seen as an add-on extrinsic to “the text’. It is no more part of the
law than the explanations and advice a client might receive from his or her lawyer.
This way of looking at law and legislation rests on, and supports, a number of
positivist assumptions: (1) law consists of general rules, (2) law achieves a material
existence because it is embodied in texts, and (3) there is a sharp distinction between
law and not-law, which corresponds to the distinction between text and context. On
the basis of these assumptions, both drafters and interpreters have become accus-
tomed to the following equation:
[law] = [rule] = [the text in which the rule is declared] = [the meaning of the
text as understood by the legislature] = [the meaning of the text as understood
by the subject] = [the meaning of the text as understood by the official inter-
preter]
This equation is the basis for reconciling the current practices of drafters and inter-
preters with democracy and the rule of law. The law inheres in the text and the text is
the same for everyone because meaning is the same for everyone.
124
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 47
Recent research by linguists and cognitive psychologists establishes that law can-
not “inhere” in a text and that meaning is not the same for everyone-meaning de-
rives from a particular context. A text is the sensory input that an individual takes in;
its meaning is constructed by drawing inferences based on knowledge stored in the
individual’s memory, specifically knowledge that appears to be relevant having regard
to the sensory input and the circumstances in question.”
Notice that the implication of this research is not that texts are meaningless, but
rather that their meaning is incomplete. Although some texts are more constraining
than others, every text has to be completed through numerous inferences that depend
on knowledge not contained in the text. This is a modest claim, but if accepted it is
enough to do serious harm to positivism. If meaning depends on the variable and un-
stable contexts that different readers (or different semiotic groups) bring to the text,
then it is impossible to draw a bright line between text and context. The content of
rules is shown to be variable and unstable, which is inconsistent with the very notion
of a rule.’
The online publication of statutes promises to be equally destructive of positivist
assumptions. The illusion of materiality created by ink on paper that is bound in
heavy books is absent in electronic publication, which emphasizes the temporal,
ephemeral, and subjective nature of knowledge. The impression of formality and spe-
cial occasion created by a statute’s distinctive appearance disappears in electronic
publication, where everything is jumbled together. Online statutes do not look very
different from other bits of information on the Web. As features like hyperlinks,
sound, and video are added, it becomes ever more impossible to distinguish text from
context.
Even if the online publication of legislation tries to conserve a recognizable text,
users will bring to it the same expectations they bring to other online information.
Online users already expect to be able to interact with a text and manipulate it by
means of hyperlinks and other technologies so as to adjust it to their own interests and
needs. They interpret this material non-textually. As the public becomes accustomed
to new technologies, it will be increasingly difficult for legislatures to rely on a text-
based vision of reality.”
” This account is drawn from D. Sperber & D. Wilson, Relevance: Communication and Cognition,
2d ed. (Oxford: Blackwell, 1995) which reviews and synthesizes competing approaches to communi-
cation theory.
‘ Critical legal scholars have made a similar claim on many occasions; it is a major tenet of their
critique. However, my analysis is based on empirical research by psycholinguists and communication
experts and potentially is demonstrable in a court of law.
47See D. Howes, “e-Legislation: Law-Making in the Digital Age” (2001) 47 McGill LJ. 39.
2001]
R. SULLIVAN – THE PROMISE OF PLAIN LANGUAGE DRAFTING
125
While plain language drafting does not require interpreters to abjure their positi-
vist beliefs, it does create occasions of doubt. For example, when traditional legisla-
tion is rewritten in plain language, interpreters are told that the words of the text have
changed, but the law is the same. This directive presupposes that law is different from
the text, that law is what is communicated by the text rather than the text itself. When
legislation is published electronically, the boundaries of text start to disappear. On the
Internet, everything is linked to everything else.
Detaching law from the text and blurring the boundaries of text lead to new ways
of understanding both text and law; this suggests new equations. The text is no longer
the law, but that which communicates the law. Accordingly, the law becomes that
which is communicated, a disturbingly chameleon conception. These new equations
have some noteworthy implications.
For example, if text is that-which-communicates, then it must include techniques
like notes and examples, for they are relied on at least as much as numbered sections
to communicate the law. There is no principled basis on which to identify the text
with numbered sections while excluding headings or examples, and there are reasons
why any attempt to do so would be unacceptable. The most important reason is fair-
ness. The plain language drafter adopts new techniques in order to communicate the
law directly to those whose lives and interests are affected. These techniques appear in
published copies of the legislation and subjects are invited to rely on them. In my
view, given this invitation to rely, official interpreters are estopped from suggesting
that non-conventional forms of communication presented as part of the statute are
context rather than text-that they do not declare the law, but merely gloss it. The
whole point of plain language drafting is to try to bridge the gap between insiders and
outsiders by writing for the outsiders. To deem that the techniques designed especially
for this audience are less reliable communicators of the law than the numbered provi-
sions is like removing the bridge to the fortress when the outsiders are halfway across.
that-which-is-
communicated. On this understanding, law becomes ephemeral and difficult to locate.
When we refer to the content of a communication, do we mean a pre-existing notion
in the mind of the communicator? Or do we mean the message that a person in fact
receives as a result of interacting with the text? Clearly, there are problems with both
views. The only certainty is that law is not a thing. It is not housed inside the statute
book and taken out for a public showing when the book is opened; it is not literally
embodied in the text. Law is best understood, or at least most accurately understood,
as a relationship rather than a thing. It is a relationship initiated by individuals at mo-
ments of application-including self-application–of the text.
When text becomes that-which-communicates,
law becomes
Plain language drafting is fundamentally a conservative effort, in the sense that it
seeks to shore up and conserve our current conceptions of democracy and rule of law.
It does not set out to undermine positivism and in truth it is unlikely to alter our basic
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGILL
[Vol. 47
understanding of law and its relation to legislative texts any time soon. It does, how-
ever, invite a different way of looking at this relationship. Specifically, it draws the
attention of drafters and interpreters to the malleable character of the text and the
make-it-up-as-you-go-along character of law.
C. Remaining Issues
I have argued that if plain language drafters are serious about enhancing democ-
racy and promoting the rule of law, they must write for the audience that has the best
claim to their assistance, which is not necessarily the audience that is most likely to
read what they have written. I have argued that the chief virtues of plain language
drafting lie in the messages they send to official interpreters about the nature of law
and the proper way to interpret statute law.
Many other issues raised by plain language drafting have not been addressed in
this paper. For example, how should the government deal with pilot projects like the
EIA rewrite? Should they become the prototype for a new format and set of drafting
conventions so that all legislation will come to resemble the prototype? Or is the idea
to do away with fixed formats and conventions and leave the drafter free to develop
different texts that respond to the purposes and needs of particular audiences?
The expectation of many plain language drafters is that once they have discovered
the techniques that facilitate direct communication with the general public, they will
use these techniques to establish new conventions and a new format and style. In my
view, this expectation is unrealistic. If effective communication is the goal, there are
no universals and endless adaptation is unavoidable. Statutes that confer benefits on
vulnerable groups in society must be drafted differently from statutes that deal with
corporate tax. Codes of conduct for specialists must be drafted differently from stat-
utes like highway codes. Monster acts like the Criminal Code” or the Income Tax Act”9
should probably be reorganized into a series of smaller acts that can be written for dif-
ferent audiences. And all acts must undergo frequent revision. What works for a given
audience or subject matter is unlikely to remain static. As communication technolo-
gies change and evolve, as audiences develop new expectations, drafting will have to
change and evolve in tandem.
In my view, plain language drafters must resist the temptation to turn the walled
fortress into a public housing project in which every unit is the same. A commitment
to direct and effective communication entails constant experiment and change in the
service of maximum personalization.
,8R.S.C. 1985, c. C-46.
49R.S.C. 1985 (5th Supp.), c. 1.
2001]
R. SULLIVAN – THE PROMISE OF PLAIN LANGUAGE DRAFTING
127
Another unresolved issue is how plain language drafting relates to the growing
harmonization of law across national borders and with international law. Harmoniza-
tion is urged on legislatures not only in the service of global markets and the global
economy, but equally to encourage the recognition of universal human rights. I have
presented plain language drafting as a desirable law reform because it acknowledges
that the general public consists of people with variable capacities, interests, needs, and
assumptions, and it attempts to accommodate these differences using evolving tech-
nologies.
Globalization, on the other hand, appears to move in the opposite direction, to-
ward greater abstraction and an indifference to local community. Instead of encour-
aging official interpreters to adopt the perspective of their local audience, it directs
them to look to the interpretations of other countries or of international courts and
agencies. Its strategy for dealing with difference is to obliterate it over time so that
eventually we may all live in Esperanto heaven and constitute a single semiotic group.
This tension between plain language drafting and globalization illustrates the ten-
sion being played out on the Internet between preserving local difference and forging
a global identity. On the one hand, the Internet encourages development of a global
language and culture, yet at the same time it has fostered the creation of new localities
and new semiotic groups. There is really no predicting what may emerge from this
dialectic.
Conclusion
If the account of plain language drafting offered here is accurate-admittedly it is
somewhat idealized-there is a good chance that it will end by destroying the statute
book as we now know it. In its place we will have something new, something less sta-
ble and homogenous, something less easy to delimit and define. Instead of a book we
will have “documents”. Here is how Schriver describes the transition from books to
documents:
Computers, consumer electronics, and technologies for multimedia are radi-
cally modifying definitions of documents and books. Document designers who
have spent their careers crafting documents such as hardback textbooks are
finding themselves challenged by advances that allow them to break conven-
tions and cross genre lines to design hybrid documents (for example, computer-
based learning environments in which a traditional book may or may not play a
part).
No longer constrained to static and linear formats, document designers can
now employ hypertext technologies to design “information landscapes” for
practical use. And with the growing affordability of integrating text, animated
MCGILL LAW JOURNAL / REVUE DE DROIT DE MCGLL
[Vol. 47
images, and digital sound, it appears that within this decade, most consumers
will get hands-on experience with documents that have a decidedly different
look and feel than those they grew up with. Such developments … put profes-
sionals in the auspicious position of being able to reinvent themselves and their
documents, leaving the word “document” as a placeholder for a text-like arti-
fact composed in print or in mixed media!’
Plain language drafters are the document designers of legislation. They currently
find themselves, in Schriver’s words, “in the auspicious position of being able to rein-
vent both themselves and their documents.” This is an occasion to be seized. I am not
reckless enough to imagine that plain language drafting will actually achieve direct
and effective communication with the subjects of legislation-not in my lifetime. But
the goal is worth pursuing and a number of benefits are likely to flow from the at-
tempt.
First, the plain language projects undertaken in Canada and elsewhere engender
research and discussion that contribute to the emerging discipline of legislative draft-
ing. If nothing else, plain language drafting is valuable because it recognizes that the
design of a legislative document has important political and legal implications. It is
not just a matter of fooling around with the form.
Second, plain language drafting makes legislation easier to read and use-both
for professional readers and those who wish to bypass the professionals and go di-
rectly to the statute itself. Testing to date suggests that most people find it better-
easier, less frustrating-to work with plain language legislation. And whether or not
they prefer it, people working with plain language legislation tend to make fewer mis-
takes.’
Finally, plain language drafting is an invitation to the official interpreters of leg-
islation to adjust their view of what they do. By undermining the simple equation of
[law] = [rule] = [text] = [meaning of the text], it draws attention to the improvisational
character of law. By attempting to communicate directly with the persons affected by
the legislation, it encourages interpreters to adopt an approach that favours the most
appropriate audience. These modest improvements are obviously not sufficient to
produce instant democracy or genuine rule of law, but they draw attention to the
problem; they make a start.
Schriver, Dynamics, supra note 37 at 4-5.
5, For a summary of the practical benefits of plain language drafting, see J. Kimble, “Answering the
Critics of Plain Language” (1994-95) 5 Scribes J. of Legal Writing 51.
