Authoring a PhD

5.A
First-order heading
5.A.i
Second-order heading
5.A.ii
Another second-order heading
5.A.ii.a
Third-order subheading
5.A.ii.a
Another third-order subheading
5.1
First-order heading
5.1.1
Second-order heading
5.1.2
Another second-order heading
5.1.2.1
Third-order subheading
5.1.2.2
Another third-order subheading
Alternatively in humanities subjects the same effect is often
achieved by mixed-together different letter and number sequences
such as this:

---

In both these examples the number sequence is overdone 
and looks ugly and hard to follow. Extending it to fourth-order
subheadings includes five or more numbers (such as 5.1.2.1.3,
which occurs in some cases): this step sends a very clear signal
to readers that you care little or nothing about the accessibility
of your text. Readers will find it difficult to tell whereabouts
they are in such an overcomplex hierarchy of headings, espe-
cially where the headings at different levels look very similar (as
in my examples above). Adopting such a schema cannot give
cohesion to an argument that has become much too fragmented.
Nor can it impart genuine order and hierarchy when an author
has not clarified her ideas sufficiently to organize her text in a
more considerate manner.
It may also be that authors who adopt complex numbering
schemas are actively encouraged by the availability of this device
to chop their argument up into ever smaller pieces. Typically
they may overdevelop an ‘analytic’ argument so as to create a
‘fruit cocktail’ effect, discussed above (on p. 70). They place so
much reliance upon the chaining of numbers or symbols at the
start of each subsection that their basic intellectual approach
alters. They start making too many distinctions, in a kind of
‘logic-chopping’ manner. For this reason my personal practice
has always been to recommend people to number only the main
sections of chapters (such as 3.1 or 3.2); and to avoid using
headings with more numbers in them (like 3.1.2 or still worse
3.1.2.1). Using numbered headings only for chapter main 
sections but not for smaller subsections seems to work best for
the vast majority of humanities and social sciences PhD theses.
Take a flexible approach to this rule of thumb, however. In
the humanities especially, you may want to try and do with-
out
any
numbered sections, if other professional writings in
your discipline have a very literary or understated feel. Here
you would rely only on the differing font sizes, emphasis and
location of various orders of headings to give a clear sense of
their hierarchy to readers. At the other end of the spectrum, if
your discipline has a strong ‘technical writing’ style, as some
areas of the social sciences do, you may wish to use numbered
second-order headings, for subsections within the main chapter
sections (that is, numbers like 3.2.2). But it is wise to hold the
line here and not to introduce four- or five-number headings
8 2
◆
A U T H O R I N G A P H D

---

(like 3.1.2.3 or 3.2.3.2.3) for smaller subsections, which will
tend to encourage you to use overfragmented modes of exposi-
tion. It is also worth remembering that across most disciplines
it will be much easier to get thesis material published as a jour-
nal paper (or even as a book), the less it seems like a report and
the more accessible the text appears. Converting an overnum-
bered chapter into a paper is not a trivial task. If you have relied
on the numbering scheme to give coherence, then you may
have to redo all the links from one section to another, and
much of the internal signposting in the chapter from scratch, if
it is to work as a paper.
(iii) The final common problem with headings occurs when
thesis authors do not use the same system of headings across
all chapters, but employ different systems at various points.
Most inconsistency problems occur because students write up
their chapters one at a time, often beginning with a typical lit-
erature review which goes over length and becomes difficult
to organize. As they write later chapters so they change their
ideas about sections and headings, and start using different
schemas, without going back to their earlier work and redoing
the headings in the new format. Whatever scheme of headings
you arrive at, it must be applied to give the same ‘look and feel’
throughout.
However, this requirement is quite consistent with the need
for your scheme to be flexibly handled, in a way that responds
to the nature of each different chapter and section, rather than
being implemented in a mechanical or robotic-looking fashion.
The system of headings stays the same throughout the main
text, but some chapters may not need to use all the elements of
the schema. For instance, you might use only first- and second-
order headings in shorter chapters, with brief sections. But then
you can introduce third-order headings in bigger chapters
which have longer sections or which handle more complex
material.
Just as a constantly updated rolling synopsis is a useful plan-
ning and revising tool, keeping you in touch with what the cen-
tral argument of your research is really about, so it can be very
helpful to maintain an ‘extended contents page’ showing the
current sequence of materials in your thesis. This page may
never be included in the final thesis, or used by anyone but
O R G A N I Z I N G A C H A P T E R O R PA P E R
◆
8 3

---

you. Instead its role is to help your planning and your orien-
tation thinking by displaying a synoptic view of how your 
thesis is organized down to your lowest order of headings and
sections. Some authors find it helpful for their extended con-
tents page to include headings and subheads and any number-
ing used, in the same font and layout as they are shown in 
the chapters, which may spread the material out over several 
A4 sheets. Others like to use a more condensed format for the
extended contents page, showing differences of emphasis, but
in more compressed ways. By keeping the extended contents
page on at most a couple of sheets of paper this approach may
give an easier overview of the structure of your material.
Devising headings and subheadings
The best way to inform your reader is to tell them
what they are likely to want to know – no more
and no less.