Authoring a PhD

Finishing a chapter
You should mark the end of the chapter by a Conclusions 
section which is at least two paragraphs long. It should have 
a heading displayed in a font which makes clear that it is not 
O R G A N I Z I N G A C H A P T E R O R PA P E R
◆
9 7

---

a first-order section. The first paragraph (or part) of the Con-
clusions should gather up the key points previously pulled out
in each of the final paragraphs for each section, and re-present
them so as to draw together the end points of each section. It is
worth writing the opening sentence of the Conclusions care-
fully, preferably in a general way which clearly breaks away
from the ending of the last section and instead encourages
readers to look back across the chapter as a whole and to assess
what they have learnt.
The second paragraph (or second part) of the Conclusions
should ‘open out’ to briefly consider one or two broader issues
raised. It should always end by establishing a forward link of
some kind to the next chapter. With a descriptive sequence of
chapters the link will normally be easy to make – for instance,
in a historical or narrative sequence, what happened next? And
in a ‘guidebook’ pattern, what links A to B? Where the chapters
discuss a sequence of analytic or argumentative topics the link
across will usually take the form of pointing to some open
issues raised by this chapter, one of which the next chapter will
address. Sometimes there are more tricky transitions, when 
a series of connected chapters ends and you have to link for-
ward to a new grouping of chapters. In these circumstances you
may want to leave a couple of blank lines to indicate that the
conclusions for this chapter alone have finished, and that some
more general comments follow. Then write a separate para-
graph or two just of linking text, drawing the connected chap-
ters together and possibly referring back to your opening
chapter plan and the sequence outlined there.
Conclusions
In the UK’s difficult and lengthy driving test there is a much-
dreaded element called the ‘emergency stop’. At the beginning
of the test your examiner tells you that at a certain random
point she will tap on the dashboard of the car with her folder,
as a signal that you must bring the car to a halt as quickly as
you can, under control and safely. Then the test starts and you
drive off, usually quite quickly forgetting about this whole idea
under the stresses and strains of negotiating traffic. Later on, as
9 8
◆
A U T H O R I N G A P H D

---

you are driving down some less populated section of road you
suddenly notice your examiner apparently having a fit and lash-
ing at the dashboard with her folder. As belated recognition
dawns, you respond by bringing your car to a screeching stop
amidst a copious cloud of burnt rubber from the tyres. For authors
of doctoral theses (and indeed other professional works) it is 
a good idea to think of an analogous emergency stop test for 
your text.
Suppose that at some random, unannounced point I take the
text away from someone who is reading your chapter. I ask her
to explain (without looking at it again) whereabouts she is in
the chapter, and what it is all about. If the text is adequately
and appropriately organized then the reader should be able to
respond:
The chapter is about the four themes W, X, Y and Z and it
has three sections. The first was about W (specifically
subtopics w
1
, w
2
and w
3
). When the text was taken away I
was in the middle of the second section covering X, having
already absorbed subtopics x
1
and x
2
. I believe that three
more subtopics x
3
, x
4
and x
5
would be handled later on in
that section. I have a clear but general idea of the topics yet
to come in the bit of the chapter I haven’t yet read, namely
that this third section will cover Y and Z together, and in 
a briefer way than the treatment of W and X.
If our mythical reader cannot respond as precisely as this,
then the chapter is too weakly structured. The worst case result
for an underorganized chapter would be if the reader responds
to the emergency stop test by saying:
I have no real clue what the chapter as a whole is about,
because the title is very vague or formalistic. From what
the author says at the start perhaps the focus is on some X
and W themes in some way? The chapter just started out
on a magical mystery tour, and has so many [or so few]
headings that I cannot really say how it is subdivided. 
I can only tell you roughly where I have been up to the
point where the text was taken away. And I have little idea
of what was to come in the rest of the section where I was
O R G A N I Z I N G A C H A P T E R O R PA P E R
◆
9 9

---

stopped, and no idea at all what remains to be discussed
in later sections. Every other page I turn throws up a new
element or a new direction in an unpredictable manner.
While it is important always to adequately organize your
text, how you chunk up your chapters must also depend a great
deal on the material that you are handling. The advice in this
chapter should not be read as a series of remedies to be mechan-
ically applied to produce chapters which are all the same.
Although chapters should generally average 10,000 words in
length, with main sections every 2500 words, that does not
mean that every chapter should have the same four main sec-
tions as every other. It is important to adjust your structures
sensitively to the material you are handling, rather than to pro-
duce robotic-looking work. An excessively mechanical applica-
tion of these (or any other) rules could mean that you subdivide
and signpost text more than you need to, producing fake sub-
sectioning and a text that is very boring for readers to plough
through.
So you need to be flexible, tuning and adjusting the principles
set out here so as to accommodate different lengths of chapters
and sections, and different kinds of material across them.
Chapters smaller than 10,000 words may need only two or three
sections, while longer ones might need perhaps five sections or
at most six sections (but not more than this). Main sections in
long chapters may need to be well organized in subsections that
are explicitly signposted, producing perhaps twelve or more
first- and second-order subheads in all.
The text box below shows a flexibly applied structure for a
middle-sized chapter (let’s say, chapter 2), with each of the head-
ings shown in its appropriate font, appearance and location.
There are three main sections, plus a short (untitled) introduc-
tion and a brief conclusions bit. The box also notes where start
and finish elements need to be more carefully written. In this
plan section 2.1 has two subsections (each with second-order
subheads), but section 2.3 is shorter and does not use any sub-
sections. And although the larger piece of text in section 2.2 is
subdivided, it is differently handled because of the nature of the
material there, using three lighter-touch groupings of paragraphs
denoted by only third-order subheads. Figure 4.1 on p. 102 shows
1 0 0
◆
A U T H O R I N G A P H D

---

the same structural information as the text box below, but in 
a more diagrammatic form. It illustrates the general point that
having a clearly recognizable and standard set of headings
across the thesis as a whole is perfectly compatible with having
chapter structures which flexibly adapt to the demands of
organizing different kinds of text.
O R G A N I Z I N G A C H A P T E R O R PA P E R
◆
1 0 1