Authoring a PhD

Analytic approach
(c)
Argumentative approach
(a)
Descriptive approach 
(for example,
start from door and go clockwise
)
Good points
(e.g. space, storage)
Basic size
,
shape
, etc.
Soft furnishings
(e.g. sofa, cushions,
books, curtains)
Door
Sofa
Cushion
Notice board
Painting
Books
Figure 3.5
Three ways of viewing my home study

---

Bad points
(e.g. eclectic furniture,
clutter overload)
Services
 (e.g. light,
windows, electricals)
Hard furnishings
(e.g. shelves, filing cabinets)
Filing cabinets
PC
Desk
Chair
Box files
Printer
More  books
Notebook PC
Window

---

become very long-winded and hard to organize mentally. These
problems may perhaps lead you to think that I am setting up 
a straw person to knock down here, and that in practice 
in advanced humanities and social science research it would 
be very hard to find people utilizing this kind of pattern of
explanation.
Think again. Most theses in these disciplines still follow a
descriptive approach, in the sense defined above, in that their 
fundamental organization is set externally to the author, by the
way things are arranged in the ‘real world’. Key forms of a
descriptive approach are:
◆
Narrative theses, which follow the pattern of a storyline set
by an external work or by another author – for example, a
critical exposition discussing Act 1 of a play, followed by 
Act 2, then Act 3, etc. This pattern is popular in literature
studies.
◆
Chronological theses, which essentially let a historical
sequence dictate their structure, beginning at the beginning
and going on until they come to the end. This pattern is
prevalent in historical studies and related fields.
◆
Institutional theses or those with a ‘guidebook’ structure –
which replicate the pattern of an organization chart, or the
relationships among different institutions, or the structure
of a piece of legislation or a set of regulations, in order to
trace out its working in loving detail. This pattern is popular
in law, public administration, social policy, and so on.
Other descriptive patterns can be envisaged, for instance spa-
tially organized work in geographical studies. Less commonly
found at whole-thesis level is the most popular descriptive pat-
tern in masters level or undergraduate essays, the ‘random
sequence of authors’. Here the order in which sources are dis-
cussed, and their relative weighting in the essay, are both deter-
mined by which sources students were able to access in the
library that week in the time available to do the essay. So a lot
is written about sources which were accessed first; rather less
about sources which the student only had a short time to
absorb; and least of all is said about sources which the student
is only pretending to have read. But even the ‘random sequence
of authors’ pattern often recurs over sections of a thesis.
6 6
◆
A U T H O R I N G A P H D

---

In lower level university studies with exams as a key assessment
method, and even for taught courses at PhD level, adopting a
descriptive approach to organizing your ideas and sequencing
your work is a popular but very damaging habit. It is prevalent
because it seems a lazy way. You just pick up an already ‘given’ or
perhaps ‘obvious’ structure existing ‘out there’, and organize your
work around it. It is damaging because a descriptive approach
demands a very high load of facts or other materials to make it
work well, and yet it often becomes hard for authors to control
and hence ends up looking very disorganized. Just as the things
which sit next to each other in my study form an eclectic list,
hard for readers to follow or understand, so things which sit next
to each other in historical time or institutional space may be all
jumbled up thematically or analytically.
But in ‘big book’ theses these difficulties are greatly amelio-
rated. The space and time constraints of lower level university
studies are not so pressing at PhD level – indeed they may not
seem to be present at all to beginning students. At doctoral level
descriptive explanations can work better, because you can assem-
ble the mass of facts and evidence needed to make the approach
look comprehensive and non-naïve. In addition, some kinds of
descriptive (externally structured) explanation are clearly popular
with and accessible to a wide range of readers, especially histori-
cal and narrative writing. The most chronological of all A to Z
storylines are biographies, which sell very widely.
Yet to make a descriptive structure work in most of the human-
ities and ‘soft’ social sciences in fact demands very high level
authoring skills. In very subtle ways you need to first articulate
and then weave into your meta-level descriptive account either
analytic concepts or argumentative themes. This thematization
of what seem to be just narrative, chronologies or ‘guidebook’
texts is an art that is harder than it looks. If you have not reached
this high level of attainment then you should always examine
carefully the three alternative approaches below before conclud-
ing that you can successfully make a descriptive structure work in
your thesis. The danger is that your thesis argument flounders 
in a disorganized fashion, presenting a jumble of complexities 
in which a single not very important feature (like temporal 
proximity in historical accounts or institutional connected-
ness in guidebook arguments) is prioritized over everything else.
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
◆
6 7

---

Such theses can often seem to be structured by no clear internal
or intellectual pattern of organization.
Analytic explanations
It is not difficult to break up and reorganize a complex description
into more analytic headings. The key step is to use organizing cat-
egories conjured out of your own brain rather than a sequence of
ideas given to you externally. For instance, an analytic approach
to describing my home study is shown in Figure 3.5(b), where 
I might structure my account around the following headings:
◆
the physical size, shape and features of the room (basically
rectangular, with a little add-on bay window);
◆
the services in the room (the windows, ventilation, lights,
central heating, plug points, etc.);
◆
the hard or fixed furnishings (shelves, bookcases,
immovable heavy filing cabinets, etc.); and
◆
the soft or variable furnishings (curtains, carpets, movable
furniture, PCs and electronic gear, books, CDs, etc.).
These different categories do not sit out there in the ‘real world’
for me to pick up ready-to-use: instead they are mental cate-
gories of my own choosing. But on the other hand they are not
rocket science and they did not take ages to devise. I hope that
these distinctions would not need a lot of explanation to be
accepted as useful and reasonably familiar by most readers. But
if I now run over what there is to see in my home study using
these headings, I am pretty sure that most people will see this
account as much clearer, as much better organized than the
descriptive approach’s almost random sequencing. As well as
providing key principles for explaining why sets of things are
treated together, the headings also capture clearly my value-
added contribution and thus help to personalize the account.
Three main types of analytic structures are used in humani-
ties and social sciences theses:
◆
Periodized historical or narrative accounts break away from
a beginning-to-end chronology, and instead chunk up the
6 8
◆
A U T H O R I N G A P H D

---

storyline into a number of clear periods. The characteristics
of each period can then be treated more synoptically. The
crucial transitions are from one period to another. They are
separated out for focused treatment, while the more
ephemeral ebb and flow of less important events within
each period is given less emphasis.
◆
Systematic accounts disaggregate complex processes into
their component parts, as in my study example above. An
overall set of phenomena (such as a change process or an
intellectual problem) is split into different components and
each aspect is treated using appropriate concepts, theories,
methods and evidence for that category. For instance, you
could split historical processes into separate economic,
political, cultural and social changes, and develop different
models of each, as well as an account of how they
interconnect. Or you could analyse a novel or a play in
terms of characters and their interactions, or identify
different elements, myths or themes woven through a
narrative.
◆
Causal analyses go further than simply handling different
aspects under category headings. They seek to reconstruct
complex multi-causation processes by grading and sifting
how influences are patterned, weighting causes against each
other, distinguishing long-term and short-term, or necessary
and sufficient causes. Very sophisticated approaches here
may trace out a complete algorithm, an analytic model of
the processes that are being studied.
An analytic structure has many advantages, so long as the 
set of organizing categories being used is simple and robust,
picking out clearly distinguishable sets of phenomena in very
clear-cut ways. To organize a whole thesis, you need a fairly
restricted structure of big, broad concepts. Fine-grain or subtle
distinctions that take ages to explain are not suitable for this
top-level organizing task, or indeed for providing an internal
structure for chapters (see Chapter 4). Robust organizing cate-
gories should also be recognizable ideas, with which readers 
can easily connect. Both these requirements may seem to 
limit the scope for you to personalize your thesis organization.
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
◆
6 9

---

They often seem restrictive for new authors who are convinced
of the uniqueness of their individual approach. But it is per-
fectly feasible to impress clear views on your chapter plan with-
out lurching off into idiosyncrasy or impenetrable distinctions.
Once you have an analytic structure of chapters it is also
important
not
to follow through unquestioningly with a further
analytic way of carving up material inside each chapter. Do not
overdo the analysis. At its limit an ultra-analytic thesis can
resemble a fairly unique (and awful) item of British cuisine, the
canned ‘fruit cocktail’. This dish consists of different kinds of
tinned fruit (like peach, apricot, pear, apple, grapes, cherries
and so on), all cut up into small cubes and mixed together, and
then completely covered in a sugary syrup. When you eat a
mouthful of canned fruit cocktail you may know intellectually
that you are consuming different types of fruit, but the tastes
are so effectively homogenized that you will have difficulty
identifying what any given cube consists of. The analogous
danger in academic life is that you wrench apart connected
phenomena to such a detailed extent that your readers lose any
grip on how the parts connect as a whole. For instance, if you
analyse a chronological process into separate analytic compo-
nents, and then analyse each of these in turn into subcompo-
nents, readers may lose any working sense of how the processes
being described operated over time, and hence find no clear
narrative storyline at all. Overextended analytic arguments can
also produce very formalistic patterns of organizing material,
with multi-layered typologies or sets of categories being
expounded which are very remote from ‘ordinary knowledge’
ways of looking at problems. In some technical or highly theo-
retical areas very formalized treatments may be acceptable,
even expected, especially in the parts of social sciences and 
philosophy. But outside these areas, they can easily look 
off-putting or impenetrable, especially where an author uses 
unfamiliar organizing concepts.