The essay below was published in The Judicial Review: Journal
of the Judicial Commission of New South Wales, 7 (September 2004): 39-56.
Spelling and punctuation reflect Australian usage.
The Architecture of Argument
Professor James C Raymond
The question of how best to structure judgments is an
ongoing one. This article is based on the premise that
judges should convey their reasoning in a form that reflects
the simple and repetitive logic of the law. It provides a
seven-step recipe for writing clearly structured judgments
that convey logical reasoning and contain context before
details, clearly partitioned issues and succinct arguments.
I once had the following exchange with a gracious judge who
allowed me to review his work in a tutorial session:
"I had trouble figuring out what's going on in this case
until I got to page 15," I said. "This is where you get
around to mentioning the issues."
"Yes, professor, I can see that."
"And now that I know what the issues are, it seems to me
that probably twelve of the first fifteen pages could be
omitted, since they have nothing to do with any of the
"Yes, professor, I agree."
"Just out of curiosity, why did you wait until page 15 to
enunciate the issues?"
"Well professor, to tell the truth, I didn't know what the
issues were myself until I got to page 15."
It was an instructive admission. Writing is often a means of
discovering what we think. It is not unusual for judges and
lawyers to discover the case as they write it.
They make a mistake, however, when they require their
readers to wander through the same process of discovery - to
follow them down blind alleys, wrong turns, false starts,
and irrelevant facts until the issues finally pop up like
mushrooms after rain.
I. The universal logic of the law
Every legal argument can be distilled to the same simple
structure, a variation of the classic categorical syllogism:
These facts (narrate facts)...
viewed in the context of this
law/contract/regulation/precedent/section of the
Constitution/principle of equity (choose one)...
lead to this conclusion (relief sought).
The logic never varies. At trial the judge's job is to
discover this pattern of thought in the morass of facts,
distortions, outright lies, genuine issues and spurious
arguments that the contending parties allege. The attorney's
job is to assist the judge in reducing the facts and
evidence to this pattern.
In jurisprudence, only four arguments can occur:
- The litigants may contest factual allegations.
- Or they may claim that the other side has cited the wrong
- Or they may argue that although the other side has cited
the right law, they have misinterpreted it.
- Or they can agree about the facts and the law, but
disagree about how one applies to the other.
Every case boils down to some combination of these four
basic disputes. There are no others. Litigants may argue
about things outside the law: technicalities in accounting
procedures, similarities among patented products, or the
reliability of laboratory tests. But these arguments involve
other disciplines. They are not legal arguments.
Even when some procedural issue is argued (venue, for
example, or timeliness), the argument will always be the
same. One side will allege certain facts in the context of a
controlling law, principle or standard, and the other side
will either dispute the facts, or argue that the wrong law
has been cited, or that the right law has been
misinterpreted or misapplied.
When several issues are involved, each must be resolved with
the same logic: certain facts, considered in the context of
a particular law, lead to an ineluctable conclusion.
The logic of jurisprudence is the same in trial courts and
courts of appeal. The only difference is that at trial,
litigants are likely to argue about both facts and law,
whereas in courts of appeal arguments tend to focus on the
law - the appellant arguing that the court below has applied
the wrong law, or misinterpreted or misapplied the right
one. Appellate courts are not equipped to examine the
evidence itself. They cannot call witnesses, examine exhibits or indulge litigants
in the sort of lengthy, unpredictable, and often disorderly
proceedings that characterise a trial. Courts of appeal may
hear arguments about the admissibility or sufficiency of
certain evidence, but except in rare circumstances they will
not second-guess trial courts on the inferences of fact
drawn from whatever evidence they deem admissible.
Because the pattern of legal logic is always the same, the
structure of an effective judgment at any level is identical
to the structure of a good brief. These genres have
different audiences, but the same purpose: to persuade.
There is one important difference. A judgment has the
advantage of authority. A judge can issue an order instead
of merely asking for one.
II. A universal outline for judgments
If the logic of the law is so simple and repetitive, why do
judges and lawyers have so much trouble organizing what they
Because despite the appearance of logic, litigation is
always messy and uncertain. It relies on "facts" inferred
from observations that cannot be replicated, reported by
witnesses who may or may not be telling the truth or by
experts who are generally contradicted by opposing experts.
Inferences made from events described by witnesses rarely
achieve the reliability of science. Even evidence that
claims to be "scientific" can be contested by other data or
other interpretations of the same data or by arguing that
the data has been contaminated.
Nor do issues arise from the facts with a logical
inevitability. Good lawyers can find many issues in any set
of allegations, some more likely than others to benefit
their clients. Unanticipated issues and surprising facts may
arise during the trial, and sometimes on appeal.
In addition, the logic of the law often melts like a pocket
watch in a surreal painting. Analogies, which are the basis
of common law (the claim that the case at bar is essentially
like a precedent), always limp. Precedents are always
Furthermore, the language of the law is rotten with
ambiguity. Despite the best efforts of legal drafters, a
motivated reader can find more than one meaning in any text.
A word like "murder" may seem plain enough – until we have
to decide how it applies in cases of abortion or assisted
suicide. A term like "marriage" may seem plain enough -
until we have to decide when cohabitation becomes marriage,
or whether one member of a same-sex union can claim spousal
benefits on the other's insurance policy. Absolutely no word
in the law is immune from the ambiguity it might contract,
like a contagious disease, in the context of a novel set of
facts. What seems like "plain meaning" when a legal text is
drafted disappears in a swirl of indeterminacy when the text
is applied to facts the drafters did not anticipate.
Jurisprudence requires lawyers and judges to control the
chaos by conveying their reasoning in a form that seems
logical. Instead of controlling the chaos, however, they
often reproduce it, failing to identify or to partition the
issues, rambling through facts and allegations without
distinguishing the credible from the implausible, switching
from one party's version to the other's as if judges were
court reporters, reproducing the testimony instead of
analysing it. Their arguments meander, just as their thoughts must have
meandered. They produce a stream of consciousness instead of
an orderly sequence, a diary of dawning awareness instead of
an engine of logic in which a result emerges from an
application of law to fact. They forget that the goal of
jurisprudence is to pluck the essential issues, the relevant
facts and controlling laws from the maelstrom of arguments,
allegations, precedents, principles and pretensions that
rage about during a trial. It is not an easy task. But it
would be easier if judges would remember the simple logical
structure that they must identify in the resolution of every
issue in every case.
Many jurisdictions publish rules to assist lawyers in
organising their submissions. These rules generally make
excellent sense. "First, tell us what the issues are," they
seem to say, reflecting an awareness that facts have no
significance until they are placed in the context of issues.
"Then tell us what the case is about" - reflecting the
frustration of judges who have to read dozens of pages
before discovering the basic fact situation from which the
case arises. And finally, "Organise the rest of the judgment
in a logical and predictable order" - a plea from readers
who are continually surprised by what turns up next in an
Paradoxically, judges sometimes forget that their readers
want precisely the same things: context before details,
clearly partitioned issues and succinct arguments. Rules for
appellate procedure generally work just as well on both
sides of the bench, and at every level, all the way to
III. A seven-step recipe for organisation
Here is a recipe for organising a judgment in even the most
- Identify and partition the issues.
- Prepare an LOPP/FLOPP analysis for each issue.
- Arrange the analysis of issues like rooms in a shotgun
- Prepare an outline with case-specific headings.
- Write a beginning.
- Write an ending.
- Review your draft with a checklist and a friend.
1. Identify and partition the issues.
Plan the body of the judgment before settling on an
Use a stack of note cards, or half sheets of paper, or the
equivalent space on a computer screen. On each card write
the word "Issue", followed by a brief statement of any
question you will have to decide.
At trial, the issues may be either questions of fact or
questions of law. At the end of the trial you will have to
present your findings and support them with reasons enough
to satisfy the court of appeal, if not the losing party.
During the trial you may have to provide written responses
to preliminary or interlocutory motions - again, with
reasons that will survive on appeal.
At trial, the issues are any reasonable and relevant
question raised by either party. To these, some judges add
questions that either they or the court of appeal might
think ought to have been raised, even if only to mention
them as a way of anticipating what might be raised at a
When judges are responsible for finding facts, they have to
support their findings with credible reasons.
(Paradoxically, juries do not have this obligation: they
find facts without revealing their reasons.) When a jury is
responsible for finding facts, you are relieved of your
responsibility to provide reasons for these findings; but at
the same time you become responsible to guide the jury's
deliberations by composing directions that they can
understand and that will also satisfy a court of appeal's
demand for legal accuracy. Either task alone would be
difficult enough; achieving both at once is just short of
On appeal, you should have the assistance of counsel in
identifying and articulating the issues. At either level,
judges in some jurisdictions use case management procedures
to have counsel clarify the issues among themselves before
addressing the court.
Determining the issues early is essential to efficiency in
writing and economy in the result. You cannot distinguish
relevant facts and arguments from pointless digressions
until you have determined precisely what questions the court
is being asked to settle. If the issues change as the case
proceeds, prepare cards for the new ones and discard those
that become irrelevant.
Partitioning the issues is essential to the structure of
your judgment. Unless each issue is clearly separated from
the others, your judgment will seem like a vast swamp -
shapeless and devoid of direction. Dividing your judgment
into discrete issues enables you and your readers to focus
on the analysis of each one individually.
2. Prepare a LOPP/FLOPP analysis for each issue
The easiest way to organise the analysis of each issue is to
follow this pattern:
- LOPP (Losing Party's Position)
- FLOPP (Flaw in Losing Party's Position)
LOPP: Respondent contends that he had not been informed of
the penalty clause in the contract.
FLOPP: The evidence shows that both the respondent and his
attorney received the contract thirty days before signing
CONCLUSION: Therefore respondent's contention that he was
unaware of the penalty clause has no merit.
When the conclusion is obvious, it may be effective to leave
it unstated, allowing your readers to make the inevitable
inference on their own. These inferences will, of course,
become explicit in the form of findings or orders at the end
of the ruling as a whole.
Sometimes it is effective to refer
to an unstated conclusion as if it were so obvious that it
can be safely tucked away in a subordinate clause (for
example, "Because the respondent had ample time to examine
the contract before signing it..."). Understatement of
this sort can be more powerful than rhetorical excess. It
implies that any reasonable reader would agree.
Be careful about using highly charged language to
characterise the losing party's position. Charged language
is a rhetorical weapon that often backfires. It pleases
readers who agree with you in advance, but it alienates
impartial readers, and infuriates the losing party and
anyone who may be sympathetic to the losing party's point of
view. There are, of course, exceptions, when judicial
indignation is perfectly appropriate and effective. But
charged language is often a sign that an argument is based
on passion rather than law. Normally, a civil society wants
judges to rise above emotion and blatant political
preferences. People who pay attention to the courts want
reasons, not feelings nor even ideals - reasons that seem
firmly grounded in law. Express the losing party's position
as effectively as you can - as if you were representing that
party yourself - and then identify the flaw in that position
with surgical detachment. If you cannot find the flaw in
your best statement of the losing party's position, you may
need to reconsider your conclusion.
Although the final logic in a judgment always resembles a
categorical syllogism (controlling law/relevant
facts/conclusion), actual courtroom argument is dialectical:
one party argues X, the other argues Y. Lawyers are always
responding to the opposing party's position. This dialectic
should be easy to find in the analysis of each issue:
One party says X.
The other party says Y.
The court says X (or Y, or possibly Z).
The LOPP/FLOPP pattern suggested earlier captures this
dialectic. But because the court's position is essentially
identical with the prevailing party's position, it is often
possible to skip one of these steps:
LOPP: One party says X.
FLOPP: But the court says Y because...
There is no reason to say what the winning party has argued,
since the court has adopted that position as its own.
Although the LOPP/FLOPP pattern generally works, there are a
One exception occurs when the controlling law is a principle
of equity or a matter of judicial discretion that must be
exercised without clear and definitive standards. In
determining custody, for example, or visitation rights,
family court judges can help calm raging emotions by
downplaying the notion of a "losing" party. Divide the
judgment into factors (for example, "Proximity to schools",
"Access to extended family" or "The child's safety"). Under
each heading, simply compare and contrast conditions at
mom's house with conditions at dad's house. A simple
objective description will suffice; often the inferences
will be obvious.
An adverse ruling in family court is never easy to accept;
but disappointed parents will find it easier to respect a
decision that focuses on the child's best interest rather
than on a finding that either party has been found a less
competent parent. Even when the decision is actually based
on the unsuitability of one parent, it does no harm for the
record to acknowledge whatever parental strengths can be
attributed to that parent along with the weaknesses that are
critical to the decision.
Another exception to the LOPP/FLOPP pattern occurs when
judges are finding facts. It generally makes sense to begin
with the position of the party with the burden of proof,
whether that party loses or wins:
Plaintiff argues that the respondent's equity in the
condominium at the time of the divorce was $250,000.
Respondent, however, presented evidence that the equity was
roughly half that amount.
After carefully weighing the evidence presented by each
side, I find that... because...
In an actual judgment each of the first two sentences would
be followed by a summary of the evidence presented, and the
third sentence would be followed by an indication of why the
judge found one party's evidence more persuasive than the
This is trickier than it seems. Many trial judges believe
that by expressing reasons for findings, particularly for
findings based on credibility of witnesses, they invite the
court of appeal to second guess them and to reach different
conclusions. On the other hand, failure to give reasons can
tempt the court of appeal to remand on grounds that the
findings were not supported by sufficient evidence. Balance
is the key. Trial judges should support their findings with
sufficient reasons to show that they are not arbitrary and
capricious. Whenever possible, they should cite specifics -
for example, evidence from documents, consistencies or
inconsistencies in testimony, conformity to or deviation
from normal human behaviour, awareness of motives for
telling the truth or for concealing it, etc. In other words,
judges can and should reveal exactly the sort of thought
processes that they tell jurors to follow in reaching a
In general, the LOPP/FLOPP pattern will help you think
clearly about the application of fact to law. It can also
protect you from your own biases. Nothing is more
frustrating to the bar and to the public than a decision
that is not supported by a clear and logical application of
law to facts. And nothing can be more damaging to public
trust in the integrity of the judiciary.
3. Arrange the analysis of issues like rooms in a shotgun
The most frequent cause of obscurity in jurisprudence on
both sides of the bench is not technical language or complex
issues or arcane subjects. It is haphazard organisation.
The easiest way to organise a judgment is to imitate the
structure of what in some parts of the United States is
called a shotgun house - a house in which each room follows
the other in a straight line leading from a front porch to a
back porch. The front porch is the introduction, the back
porch the conclusion. Each room between contains the
analysis of a particular issue.
Once you have determined the issues, arrange them in a
sequence that makes sense. If you have written each issue on
a separate card, you can spread the cards across a table and
select the sequence that works best.
Sometimes there will be threshold issues (standing, for
example, or jurisdiction or timeliness); normally these are
dealt with first.
Sometimes issues can be grouped in categories (for example,
three dealing with the admissibility of evidence, two
dealing with jury instructions, five dealing with
sentencing). Sometimes the issues can be arranged in a
logical chain, each issue dependent on the other for its
viability. Sometimes each issue is completely independent of
the others. In this situation, consider arranging the issues
chronologically, if the material allows it.
After reaching a decision on a dispositive issue, the others
generally become moot. On occasion, however, judges will
analyse these moot issues anyway, on the theory that if they
are reversed on the dispositive issue, ruling on the others
will save the litigants the trouble and expense of further
litigation. If you do this, be sure to announce your
intention in advance. Do not surprise your readers by having
them read your analysis of a dozen issues only to discover
at the end that the moving party had no standing in court.
The analysis of each issue should be self-contained, like a
stanza in a poem or a room in a shotgun house (stanza
actually means "room"). You should have as many rooms as you
In some cases, a section equivalent to a foyer needs to be
added: an antechamber just after the introduction and before
the analysis of the first issue. This section is necessary
in cases that cannot be understood without a detailed
narration of facts or a review of procedural history.
Although a "foyer" for an extended facts, background or
procedural history may be necessary at times, more often
than not it can be avoided by writing a beginning that
provides an essential overview (see step 5, below), saving
necessary details for the analysis of the issue to which
they are most relevant. Narrating the detailed facts twice -
in the beginning and in the analysis of the issues - creates
unnecessary work for yourself and your readers.
4. Prepare an outline with generic and case-specific
If a judgment is very short - two or three pages - it may
need no headings. In longer texts, headings are extremely
helpful, particularly to readers who want to read your
argument as quickly as possible. .In judgments that include
a table of contents, headings provide a road map,
foreshadowing the journey you want your reader to take.
Within the document, headings serve as signposts marking the
boundaries between various stages of the journey. They show
where the analysis of each issue ends and another begins. To
serve these functions effectively, headings mus. be as brief
as possible. They should not be entire arguments (though it
is often effective to put a brief summary of an argument
immediately after each heading).
There are two kinds of headings: generic and case specific.
Words and phrases like "Introduction", "Background",
"Order", "Cases cited", "Issues" and "Findings of fact" are
generic headings. Generic headings can be transferred from
case to case, regardless of the facts and issues.
Although generic headings are useful, even more useful are
case-specific headings - headings like "Was the warrant
valid?" or "What is the meaning of 'obscenity' in section
905?" These headings differ from generic headings in that
they are tied to the facts of a specific case. They mark
boundaries between the analysis of separate issues.
Case-specific headings enable future readers (such as
lawyers and law students) to go directly to those sections
they suspect might be helpful to other cases.
There are three ways to phrase a case-specific heading. You
can phrase it as an argument:
The University of Montevallo is not an Agency of the State.
You can phrase it as a question:
Is the University of Montevallo an Agency of the State?
Or you can phrase it as a topic:
Some judges prefer argumentative headings, never wanting to
pass up an opportunity to press their point of view. Others
think topics or questions are more effective as headings
because they convey a sense of detached objectivity, which
is in itself a persuasive stratagem. It is a matter of
personal preference, based upon the authorial persona you
want to create and on the way you think a particular reader
or set of readers is likely to react.
Even though you should write every judgment as if you expect
your readers to follow it from beginning to end, chances are
they will not. Effective headings will aid those readers who
raid your text like marauding pirates, looking for what
interests them and ignoring the rest. Make it easy for them
to find whatever they are looking for.
No matter how you phrase them, however, headings should be
clearly foreshadowed by the end of the introductory section
(see step 5 below). And they should be phrased in such a way
that they are intelligible to an educated non-lawyer who
knows nothing about the case in advance.
Here, for example, is a heading that requires far too much
knowledge of local law:
The holder of a perfected security interest is not entitled
to negate the State's rights under the statutory "warrant
hold" provision of the VIP Government Code, section 403.055
The same issue could have been stated much more clearly in
Can the state withhold Medicare funds from creditors of a
bankrupt nursing home that has failed to pay its taxes?
5. Write a beginning
It may seem odd to suggest writing an introduction at this
stage, after you have already developed the heart of your
argument. But you are not in a position to write an
introduction until you know what you are going to introduce.
Sometimes you have no idea what the issues are, or how many,
or how they should be resolved, until you have drafted a
LOPP/FLOPP analysis for each issue.
Avoid beginning with technical, dry or uncontested
assertions. Imagine, for example, the reaction of weary
readers with busy schedules when they see an opening
paragraph like this:
Pursuant to Local Patent Rule 4-5(b), Defendant National
Compuchip Corporation ("Compuchip") challenges the Claim
Construction Brief filed on February 27, 2003 by Plaintiff
Laserop, Inc ("Laserop"), on issues of claim construction
for US Patent Nos 5,944,807 ("the '807 patent") and
6,098,141 ("the '141 patent"). Exhs A and B, The Laserop
patents. Compuchip's proposed interpretation of the terms
and phrases in the claims of the '807 and '141 patents are
set forth in Compuchip's Interpretation Chart for the Claim
Terms! Phrase Recited in the Asserted Claims of US Patent No
5,944,807 ("the '807 patent") and US Patent No 6,098,141,
which is attached hereto as Exh C.
lf you are a typical reader, you probably did not read this
example in its entirety. You skipped over it as soon as your
eyes glazed over. Yet some judges are convinced that they
are bound by tradition, rules or logic to begin their
judgments with a reference to the rule that gets one party
or the other into court. There is something logical about
this convention; after all, how can we decide a case if the
litigants have not established standing and jurisdiction?
But then again, if there is no contest about standing or
jurisdiction, why waste the opening lines establishing
something that can be safely presumed?
Then, too, once they have mastered the numerical references
in a particular case or a particular statute or a particular
set of rules, some judges forget that shorthand references
are meaningless to readers who are not already intimately
familiar with the same material. References like "Local
Patent Rule 4-5(b)" and "the '807 patent" do not actually
communicate information; they merely remind a small set of
readers in the know. Granted, this beginning would make
sense to the parties involved in the case, but it wouldn't tell them anything
new. So to whom is it useful?
When jurisdiction and standing are uncontested, starting
with "Pursuant to" is like putting a hotdog stand on prime
real estate. The first paragraph and the last are possibly
the only places where you can count on the reader's
attention. Why waste this space by filling it with
uncontested assertions or with information the reader can be
presumed to know?
Similarly, imagine the reaction of their readers who
encounter opening lines like these:
Declaratory judgment (article 453 cpc)
This Court, having examined the proceedings and the
exhibits, considered the arguments of counsel, and duly
deliberated, doth now render the following Declaratory
This self-congratulatory gambit serves no purpose. It is a
sort of judicial throat clearing. It enables you to put
something on paper before getting around to the case at
hand. Why not just get around to it? Skip the throat
An effective introduction provides two things: a synopsis of
the facts and a brief statement of the issues. Begin with
what you would tell your next door neighbours if they were
curious about the case. Use ordinary, neighbourly language.
Avoid jargon. Tell a brief story indicating the human
conflict, "who did what to whom" or "who's arguing about
what". Then state the issues - the questions of fact or law
that you need to settle.
In cases destined for the highest courts, often the fate of
the particular litigants is less important than what the
decision will mean for other litigants in similar
situations. If the issues have far-reaching implications -
if, for example the suit is intended to establish or to
challenge an important public policy - you might start with
the issues and then summarise the facts.
The combination of facts and issues provides the context
that gives meaning to everything that follows. In addition,
by delineating the issues in a few lines, you can foreshadow
the structure of the argument to follow. Here is an example:
Harry Saunders was convicted of assault, battery, rape and
murder, each in the first degree. According to the evidence,
Saunders wore gloves and a mask when he committed these
crimes, concealing his identity from his victim and from
witnesses on the scene.
In this appeal, Saunders argues that the lineup in which he
was identified was suggestive, that articles of clothing
used in his identification were illegally seized from his
apartment, and that he had no access to counsel
at key points during the investigation
This beginning is exceptional not only for what it does, but
perhaps more importantly for what it does not do. It does
not establish standing or jurisdiction with the ubiquitous
phrase, "Pursuant to Rule 123 appellant asks..." It has no
legal jargon or long, tangled sentences. In fact, there is
nothing in this opening that would seem odd or technical in
a good newspaper. And that, despite whatever misgivings you
might have about the media, is an excellent standard for
The writer (a judge in Idaho) also avoided citing specific
sections of the code and specific references to precedent.
He did not feel obliged to tell us that assault, battery,
rape and murder are illegal activities (for example,
"contrary to sections w, x, y and z of the Criminal Code").
Nor did he feel obliged, at this stage, to tell us what
statutes, precedents or standards the appellant had invoked
in support of his claims. This may be essential information
at some point - the precedents will have to be cited and
distinguished, the statutes and standards may have to be
quoted if there is any dispute about their meaning or the
application to this particular set of facts. But details of
this sort should be saved for the sections in which issues
are analysed. No need cluttering the opening paragraph with
more information than the reader needs at this point.
This beginning provides the necessary context for
understanding the analysis that follows. You can even
predict the headings: "Lineup identification"; "Search and
seizure"; "Access to counsel". And in predicting the
headings, you are predicting the structure of the rest of
the document. You are, in effect, promised an easy and
interesting read. Although judges are not obliged to make
their writing interesting, doing so does have the effect of
helping the reader pay attention to the argument.
In this case, the writer felt the need to interpolate a
detailed narration (foyer) between the opening paragraphs
(the front porch) and the analysis of the first issue (the
first in a series of rooms). He did this by telling the
story of the lineup in which Mr Saunders was identified,
beginning with "There were three lineups. The first
occurred.. . The second occurred... The third occurred. . ."
In most cases a simple story-plus-issue is the best way to
gain the reader's interest and attention. But the temptation
to write abstractly is hard to resist. Here is the opening
paragraph in a case about unlawful detention:
This is an application supported by an affidavit in
which the applicant is seeking to be admitted to bail pending her
trial. The affidavit discloses that the applicant who has
been in custody since October, 1985 was on 3rd December,
1985 committed to the High Court for trial for the offence
of Infanticide. On 18th December, 1985 she applied to the
High Court at Kitwe to be admitted to bail pending her
This is an adequate beginning, but it reads like an abstract
problem in the law instead of what it really is, a case
about a young woman who has been improperly held in jail
without bail. Starting with the story would have given the
case the sense of urgency and human significance it
Rosemary Chilufya has been in jail for nearly five
months, awaiting trial on a charge of infanticide. The High
Court has refused to set bail, on the ground that
infanticide is a form of murder, and murder is not a
bailable offense. A threshold issue in this case, however,
is whether the Supreme Court has the authority to...
Stating the issues effectively requires steering a course
midway between too much detail and too little. The example
below provides too much detail - too much because it
overwhelms the reader and predicts what follows in
- The issues in this appeal in respect of the Appellant's
1994 taxation year are:
- Whether the Appellant, in determining LCT liability under
Part 1.3 of the Act, is entitled to deduct the amounts of the
Estimates from its "capital", or whether such amounts are to
be included in its "capital" :
- as "reserves" pursuant to ss 181(1) and 181.2(3)(b), or
- as "other surpluses" pursuant to s 181.2(3)(a);
- Alternatively, if the Estimates are "reserves" or "other
surpluses", whether the Appellant, in computing its income
under Part I of the Act, is entitled to deduct the amounts
of the Estimates from its revenue;
- Whether the Appellant, in determining LCT liability under
Part 1.3 of the Act, is entitled to deduct the $37,481/776
amount" as a "deferred tax debit balance" within the meaning
of s 181.2(3)(h).
The other extreme is to provide too little detail:
The issue is whether the tax returns filed by the appellant
in 1994 were accurate.
This version does not predict the structure of what follows,
nor does it give the reader a glimpse of the grounds on
which each side bases its argument.
It is also possible to provide too much and too little
detail at the same time - too much by including information
the reader does not need at the outset; too little by not
explaining what is at stake and by presuming a reader who
knows the code by heart:
The issue is whether the appellant is entitled to deductions
pursuant to ss 181(1), 181.2(3)(a), 181.2(3)(b), and
181.2(3)(h) of Part 1.3 of the Income Tax Act.
A good statement of issues foreshadows the structure of what
follows and provides the reader with a glimpse of the
grounds of the
argument. It does not cite laws, precedents or records that
can be more usefully cited in the analysis section. In this
particular case, after a brief description of what the
appellant claimed in his tax returns, the issues might have
been effectively stated like this:
The issues are:
- Whether the appellant is entitled to deduct the amounts of
the estimates from its "capital".
- Whether the appellant is entitled to deduct the amounts of
the estimates from its revenue.
- Whether the appellant is entitled to deduct the
$37,481,776 as a "deferred tax debit balance".
A good beginning makes the reader want to read more. A
notable example is this introduction in a per curiam by the
Ontario Court of Appeal:
Professor Starson is an exceptionally intelligent man.
His field of expertise is physics. Although he has no formal
qualifications in that field, he is in regular contact with
some of the leading physicists in the world. In 1991 he
co-authored an article entitled ‘Discrete Anti-Gravity’ with
Professor H Pierre Noyes, who teaches physics at Stanford
University and is the Director of the Stanford Linear
Accelerator Center. Professor Noyes has described Professor
Starson’s thinking in the field of physics as being ten years
ahead of its time.
Unfortunately, Professor Starson has a history of mental
illness, dating back to 1985. Unfortunately, Professor Starson has a
history of mental illness, dating back to 1985. He has been
diagnosed as suffering from a bipolar affective disorder. On
several occasions during the last 15 years he has spent time
in mental institutions. In November 1998 Professor Starson
was found not criminally responsible on account of mental
disorder on two counts of uttering death threats. In January
1999 the Ontario Review Board ordered that he be detained at
the Centre of Addiction and Mental Health (the Centre)".
Notice that this passage does not call attention to itself
as writing. The words are transparent, invisible, like
lenses through which we see characters and events. The
writer doesn’t seem to be trying to write. The art conceals
the artifice. It’s as if the story wrote itself. But of
course it did not. A beginning like this is carefully
crafted, a combination of talent and craft.
In this case, the plot thickens when we find out that the
unusual Professor Starson "has a history of mental illness."
And it thickens further when we discover a few sentences
later that he does not want the medication the Ontario
Review Board wants to give him, because it would cloud his
mind and hinder his ability to conduct his theoretical
A beginning like this entices the reader to continue
reading. Who would not be curious to know how the case was
6. Write an ending
Your concluding section may include only an order. However,
if you think the court above yours, or the press, or the
losing party might miss the essence of your analysis, use
your conclusion as a summation. Repeat your analysis, but in
different words, and succinctly. Brevity is essential.
The concluding section also provides an opportunity for
obiter dicta - instructions to the bar on related matters that
are not logically essential to the case you are deciding.
And when your decision is based on common sense or pure
equity, the concluding section can include what I like to
call the "To-rule-otherwise" trope. Judges rely on this
device when they have little or no law to justify their
decisions. "To rule otherwise would be to invite..." they
say, and then list the horrible, unjust, and illogical
things that would follow from a different decision.
In a very short judgment, where repeating the reasons would
be tedious, a conclusion that includes an order without
repeating the reasons may be adequate:
For the reasons above, plaintiff's Motion to Remand is
granted. This action is remanded to the Circuit Court for
Barbour County, Alabama, Clayton Division. In addition,
defendants will pay all just costs and expenses, including
attorney's fees, incurred as a result of the improper and
groundless removal of this case.
In a judgment of any complexity, however, an ending of this
sort misses an opportunity to revisit the argument. A brief
review of the analysis, like the one below, can assist the
Defendant, Tarwater Tobacco Co., has succeeded in having
this case removed from state to federal court on the ground
that Tarwater’s local agents were named as co-defendants by
plaintiff as a ruse ("fraudulent joinder") to obtain a
favorable local venue.
The standards for removal on the basis fraudulent of joinder
are quite high. In this case, Tarwater would have had to
prove either that there is no possibility of a verdict
against the local defendants, or that the complaint against
them was based on false information.
Tarwater has met neither standard. There is no evidence of
fraudulent information in the joinder. Nor is there any
question that a jury would find against Tarwater’s local
agents if the facts alleged are proved at trial.
For these reasons, the case is remanded to the Circuit Court
for Barbour County, Alabama, Clayton Division, from which it
Costs and attorney’s fees are assigned to Tarwater. Their
failure to provide credible evidence for their claim amounts
to a frivolous delaying tactic, taxing the plaintiff with
unnecessary costs and taxing the resources of this court.
It may seem paradoxical that a good ending resembles a good
beginning (which, in turn, often resembles a good head
note). The resemblance is not accidental. Your audience does
not necessarily read from top to bottom. If they get lost in
an argument, they may flip to the end, hoping to find a
synopsis there. They will not be helped by a conclusion that
says merely "For the foregoing reasons... ," sending them
right back to the thicket they had just abandoned. An
effective conclusion summarises those foregoing reasons in a
nutshell, in plain English, without repeating citations and
references that are already included in the body. Here is
how the Ontario court concluded the case about Professor
Putting aside any paternalistic instincts – and we
think that neither the Board nor the appellants have done so – we
conclude that Professor Starson understood, through the
screen of his mental illness, all aspects of the decision
whether to be treated. He understands the information
relevant to that decision and its reasonably foreseeable
consequences. He has made a decision that may cost him his
freedom and accelerate his illness. Many would agree with
the Board that it is a decision that is against his best
interests. But for Professor Starson, it is a rational
decision, and not one that reflects a lack of capacity. And
therefore it is a decision that the statute and s. 7 of the
Canadian Charter of Rights and Freedoms permit him to make.
The appeal is dismissed.
7. Review your draft with a checklist and a friend.
Persuade a friend, preferably a non-lawyer with no knowledge
of the case, to help you review your draft with the
Ask your friend to tell you, after reading only the first
page, who did what to whom and what issues need to be
- Test the overall structure by asking your friend, after
reading only the introduction, to guess what headings will
follow. If there is a good match between the introduction
and the structure that follows, your friend should be able
to guess, in substance, the case-specific headings that
separate the analysis of each issue from the others.
- Ask your friend to tell you, after reading the last full
page, what you decided and what grounds you give for the
- Ask your friend to locate the beginning and the end of the
analysis each issue, and to tell you the losing party's
argument and the flaw you found in it.
- Check for economy and consistency. If you announced five
issues at the outset, be sure that you have analysed five
issues. Delete any information that is irrelevant to the
issues. Look for repeated information; see if it can be
mentioned in one place and omitted in the other.
If your friend doesn't answer any of these questions to your
satisfaction don't explain. Revise.
A well-written judgment is as smooth as a grape. There is
nothing extra. Once you reduce it to essentials and organise
it coherently, you are ready to revise and polish for style.
IV. Recommended Reading
Garner, Bryan A, The Winning Brief: 100 Tips for Persuasive
Briefing in Trial and Appelate Court, 1999, OUP, New York.
Raymond, James C, "Legal Writing: An Obstruction to Justice"
(1978) 30 Alabama Law Review 1-17.
Raymond, James C, "
Writing to Be Read: or, Why Can't Lawyers
Write Like Katherine Mansfield?", New Zealand Law
Conference: The Law and Politics, Conference Pap . 1993, vol
2, pp 210-216. Reprinted in (1997) 3 The Judicial Review
Raymond, James C and Goldfarb, Ronald L, Clear
Understandings: A Guide to Legal Writing, 1983, Random
House, New York.
Stark, Steven D, Writing to Win, 1995, Doubleday, New York.