Authoring a PhD

Vladimir Nabokov
1
A
ny large text has to be broken up and arranged into a set of
chapters. This task may seem unproblematic. First think
about how many thousand words you want to write, and then
how many chunks of text you need to split up this total effec-
tively. Next settle on what topics to begin with, and where you
want to end up. Then fix on some way to get from alpha to
omega. So far, so straightforward. But there is a bit more to it
than that. One of Neil Young’s ironic songs has a record producer
telling a rock artist that they have a ‘perfect track’, although they
don’t yet have either a vocal or a song. ‘If we could get these
things accomplished,’ he says, ‘nothin’ else could go wrong.’
2
Planning a thesis from a blank-canvas requires a similar heroic
optimism and there are multiple considerations to keep in mind.
Your structure has to be accessible for readers. They must see
the sequence of chapters as logical, well organized and cumula-
tive. At the same time, if you are to understand what you are
about, the overall thesis plan has to sustain your progress as an
author and researcher. It must keep your argument on track,
motivate you to move on, and facilitate the development of
your methods and approach. The succession of chapters has to
be related in some definite and planned way to the timetable
for your research. The vast majority of PhD students (around
four-fifths at a guess) are ‘serial’ authors. They find it easiest to
write chapters in a single sequence, starting chapter 5 only
43

---

when chapters 1 to 4 are already pretty well defined. Generally
speaking, writing chapter 5 when all you have to look at are dis-
parate parts of (say) chapter 3 and chapter 6 is going to be a
much more difficult proposition. But on the other hand, writ-
ing up your thesis so that its chapter sequence just records what
research you did, in the order that you did it, can produce very
incoherent structures, which cut across or obstruct the current
organization of your argument and thought. Getting to a bet-
ter, 
designed
chapter structure often influences how good your
doctorate is.
In this chapter I look at three different ‘cuts’ into the prob-
lem of organizing the component parts of your thesis into a 
storyline. The first way of looking at the issue focuses on the
relationship of the whole and the core in your thesis, the 
core being the most value-added bits, the sections where you
make a distinctive contribution to scholarship or research. The
second cut looks at the choice between ‘focusing down’ or
‘opening out’ in the overall sequencing of materials. How you
sequence elements often influences the weights which you give
each component of your thesis, in terms both of text space and
of research and writing time. The way that you make these deci-
sions can affect readers’ view of your work and your own effec-
tiveness as an author and researcher. The third perspective
focuses on choosing a strategy of explanation from a limited
number of options. At the broadest level, there are actually only
four possible ways of expounding your materials in creative
non-fiction writing. Each of these options has its attendant
advantages and disadvantages.
The whole and the core
There are two things to be considered with regard
to any scheme. In the first place, ‘Is it good in
itself?’ In the second, ‘Can it be easily put into
practice?’
Jean-Jacques Rousseau
3
Anyone planning a long text needs to think logistically for a
moment. Leave aside the intellectual issues of what substantive
4 4
◆
A U T H O R I N G A P H D

---

material to write and just ponder for a bit how much, what
kind, in what order. A ‘big book’ thesis is a particularly fraught
context in which to set out to write what is good or true before
putting some numbers in the frame. In the first place universi-
ties now impose some important legal restrictions on what your
doctoral dissertation can look like. In the past many people
overwrote big book theses, greatly prolonging the time spent
on them and creating long tomes that were excessively onerous
to get examined. Nowadays any responsible university will
limit the maximum time that you can spend on a PhD – 
usually allowing from five to eight years of full-time study, but
more pro rata for part-time students. If the thought of (say) a
six-year-long project makes you shudder, as it should, do not be
fooled into thinking that this limit is purely notional. Every
year there will be people who come up to the limit and some
who overrun it.
Just as no one should go on and on as a permanent student,
so doctoral theses are now normally limited to a maximum
length, which may vary a little from one university or disci-
pline to another. In Europe and Britain where the ‘big book’
thesis remains predominant in ‘soft’ disciplines, the upper limit
can be safely thought of as 100,000 words – which is about 330
pages of A4 paper typed double-spaced. One A4 page is about
330 words, so that 1000 (or 1K) words cover three pages.
(Obviously you should check the specific regulations applying
to your discipline at your university and adjust my advice here
to fit well inside your formal limit if it is less than 100K words.)
You must take this constraint seriously from the start and make
sure that you do not overwrite it. If you work away on your
chapters in isolation, one at a time, it is very easy for hard-
working people to write 125,000 or 150,000 (even 200,000)
words of text without appreciating how the numbers are stack-
ing up. At a late stage in your research to realize that you have
25 per cent or 50 per cent more text than you need or can 
submit is a very great shock. It can take weeks or months of
painstaking work to make cuts of this magnitude in a complex
text. And cutting out whole chapters at a late stage can be
almost equally disruptive.
In fact the danger of overwriting is so acute that you need to
make sure you come in well 
within
the formal limit. A useful
P L A N N I N G A N I N T E G R A T E D T H E S I S
◆
4 5

---

general rule is to produce a main text that is no more than 
four-fifths of the permitted length. A formal words target
includes everything – all footnotes or endnotes, all appendices,
data tables, figures and diagrams. The only thing normally
excluded is the bibliography – an exhaustive alphabetical list-
ing of every book, paper, document or other source cited, which
every thesis must have in its closing pages. To be on the safe
side, therefore, write no more than 80 per cent of the permitted
number of words in your main text. An overall thesis constraint
of 100,000 words means that your main chapters should not
exceed 80,000 words. The 20,000-word difference here partly
gives you some space for the notes, appendices and other 
supplementary materials. It also includes an insurance margin
of around 10,000 words in case some of your chapters prove
stubbornly longer than planned.
In terms of what happens to your research 
after
it is finished,
a main text of 80,000 words is also a lot better. At this length
your thesis may be potentially publishable in cut-down form as
a book, while one at the legal limit will be far too big (see
Chapter 9). The average academic book is around 70,000 words
long, and the closer you write to that kind of figure the less
revising work will be entailed in converting your thesis into a
monograph. Cutting (say) 100,000 words down to this length
may not seem too difficult a task. In fact, it means losing a third
of your work, and a cut of this magnitude could take several
months work to achieve.
There are not usually formal rules about the 
minimum
length
for a doctoral thesis. But informal lower limits often do apply.
Where universities follow the ‘big book’ thesis model, then aca-
demics generally interpret regulations specifying that a doctor-
ate must make a ‘substantive contribution to knowledge’ to
mean a pretty substantial tome. The one exception is disserta-
tions using some condensed form of expression, such as math-
ematical exposition or a very formal, technical way of
expressing arguments. But in these disciplines ‘big book’ theses
are now rarely used and shorter ‘papers model’ dissertations are
anyway the norm. Another consideration is that most universi-
ties in Europe and North America have a second-tier post-
graduate thesis qualification below the PhD level, for which
candidates do not need to undertake original research and
4 6
◆
A U T H O R I N G A P H D

---

which has a lower maximum word limit. In the UK, for instance,
this non-doctoral research degree is called an M.Phil. (Master of
Philosophy) and it requires people to write a satisfactorily pre-
sented thesis of no more than 60,000 words on a worthwhile
topic. So there is a danger that PhD examiners presented with 
a short thesis of say 55,000 words may feel that it is too insub-
stantial to qualify for the doctorate, and perhaps operates more
at the M.Phil. level. Wherever such second-tier research degrees
exist, doctoral students not doing mathematical or formal work
are well advised to write more text than the upper limit for the
lower degree requirement. Your thesis should always look and
feel like a doctorate to the examiners.
Once you have set the length of your main text, ideally at
80,000 words, you need to cut it up into chunks. A basic prin-
ciple of organizing any piece of text is that it should be sub-
divided evenly, so far as possible, in this case into chapters.
Regular chunking up of text fosters consistent expectations
amongst readers: they know in advance how long chapters are.
In addition, regular divisions always look better organized and
controlled. To determine the number of subdivisions needed,
bear in mind that a chapter has a practical maximum length of
around 10,000 words. Chapters more extended than this length
make it much harder for you to organize them internally and to
control their argument effectively (see Chapter 4). Long chap-
ters are also more difficult to convert into articles in academic
journals, for which the optimum length is no more than 6000
to 8000 words. Conference papers should be even shorter,
around 5000 to 6000 words long. A 10,000-word chapter can
normally be edited down to form a decent 8000-word journal
article. With a lot of surgery it is also feasible to recast most of
it as a paper for an academic conference. But a chapter of
15,000 words will be effectively unpublishable in either form.
At this length it will need radical rewriting if it is ever to see the
light of day.
Chapters must also be of a certain minimum length if you
are to fulfil your key mission as an author and successfully
manage readers’ expectations. A short chapter, one of less than
about 6000 words, will be confusing for readers. It can easily
seem insubstantial and disappointing. It may even appear as a
‘fake’ element that you have inserted on your contents page, to
P L A N N I N G A N I N T E G R A T E D T H E S I S
◆
4 7

---

try and mask an otherwise obvious gap or unsuccessful patch in
your research effort.
Of course, theses vary a great deal in how far they can be
structured into similarly sized chunks. So these targets and lim-
its are only indicative. There will be many occasions where you
have to interpret them a bit flexibly. Yet it is a good idea to be
very sceptical about writing chapters that are much longer or
much shorter than 10,000 words. This central target length can
be pushed up or down by 2000 words either way without doing
any great harm. But chapter lengths should not go lower than
about 8000 words or higher than about 12,000 words, except
for the most pressing and exceptional reasons. Of course, it is
often hard to predict at the planning stage how long chapters
will turn out in the writing. If you end up with a substantially
oversized chapter, say one that is 17,000 words long, the best
strategy is to split it into two new, evenly sized chapters of
around 8500 words each. Do not try to struggle along trying to
organize so much text as a single unit. And do not ask your
readers to cope with following an argument at the original
monster length.
An overall text of around 80,000 words, evenly divided into
chunks averaging 10,000 (or 10K) words each, implies that your
thesis will need around eight chapters. The 8
⫻
10K format is 
a very potent one. It can usefully serve as a strong benchmark
against which you should measure any different chapter struc-
ture. With eight chapters your contents page will easily meet
the ‘seven is a magic number’ criterion (see p. 35 above). Your
readers can hold the whole sequence in the forefront of their
attention, and so can you. But if your structure has more than
10 or 11 chapters you will be unable to pay attention to it or
envision it as a whole, and you may react by randomly ‘forget-
ting’ chapters or losing track of the sequence. Again what is true
for you as author here will also be true for readers. Give them
14 chapters to keep in mind and you can be almost certain that
the overall pattern of your argument will become less visible
and harder to follow.
People often feel that the 8
⫻
10K norm is too restrictive and
that they can handle many more chapters in their thesis by
dividing it into parts, where each part is a set of connected
chapters. For instance, a 15-chapter thesis may be too complex
4 8
◆
A U T H O R I N G A P H D

---

to envision clearly, but the idea is that it could be more man-
ageable if divided into three parts of five chapters each. This use
of parts, simply to manage an inflation of the number of chap-
ters, should always be avoided. Your organizing problems will
not go away, anyway, because the individual chapters will still
become too small and fragmented. Conspicuously brief chap-
ters will seem bitty and short-weight to readers whether they
are linked together into parts or not.
A two-tier structure of parts sitting on top of chapters can
also seem attractive as a way of signalling to readers that there
are important continuities between chapters. For instance, it
might be that chapters 1 to 4 deal with different aspects of one
meta-topic, and chapters 5 to 8 are about a second, so that a
two-part division will highlight this ‘meta-structure’ for readers.
Similarly, different parts may use different methodologies, or be
focused on different levels or aspects (for instance, national
processes versus local processes). A part structure is more legiti-
mate here, and may have something to recommend it in some
circumstances. But a two-tier structure still requires careful
management. For new authors it is a complication that is often
mishandled, and so it is best avoided if possible. For instance,
you can often indicate continuities between groups of chapters
more simply by referring to the links between them in their
titles. Ideally then you should pursue a clean and uncompli-
cated 8
⫻
10K structure for your main text, without any other
organizing devices above the chapter level.
So much then for the organization of the whole. But this sec-
tion is also about the 
core
of your thesis – which may be simply
defined as all those sections with high research value-added.
The core contributes to originality either by ‘the discovery of
new facts’ or by ‘the exercise of independent critical power’.
This set of chapters contains all the most substantively new or
different sections of your research, the ones that determine if
you get a doctorate or not. In a ‘big book’ thesis not all of your 
doctorate can or should fall into the core. There will also be a
certain irreducible amount of non-core materials, composed of:
◆
Lead-in material, which introduces and sets up core material
for readers so that it is understandable and accessible.
Sometimes dismissively labelled as ‘throat-clearing’ stuff,
P L A N N I N G A N I N T E G R A T E D T H E S I S
◆
4 9

---

lead-in sections or chapters always require careful
management. None the less they often loom much larger to
students in terms of their length, and their writing and
rewriting time, than their eventual role in the final thesis
would justify. Readers often page through lead-in materials
quite quickly, looking mainly for ‘the beef’ to be found later
in the core sections.
◆
Lead-out materials do the ‘book-closing’ role for large theses,
providing an integrating summation or restatement of what
has been found, and setting it in a wider context.
When thinking about how to organize these three types of
materials (lead-in, core and lead-out), it is vital that so far as pos-
sible they should form distinct blocks in this sequence, shown
in Figure 3.1. They should not be split up and scattered around
the thesis in little chunks. Readers must be able to clearly iden-
tify the core as a set of discrete, high value-added chapters. They
should never have to search for smaller nuggets of originality
dispersed in mixed chapters that also contain other kinds of
material. The point of the lead-in materials is simply to frame,
highlight and lead up to the core. In particular, they should
ensure that readers can appreciate the originality and the use-
fulness of what you have done in your central research activities.
To get a doctorate (and to do a good thesis more broadly) the
size of the core matters a great deal. You must make sure that
there are enough core chapters, and that they are big enough 
in terms of the total wordage of your thesis, to colour the whole
thing as an original piece of work. My suggested rule of thumb
for ‘big book’ theses is that 50,000 out of the 80,000 words of
main text must be core materials. That is, appreciably more
5 0
◆
A U T H O R I N G A P H D