UNIT I: The Research Article

This first Unit introduces the idea of the Research Article. We will analyze one article from
several points of view. First, we will look at its macro-structure, at how it can be divided
into recognisable sections, each fulfilling distinct objectives. Then, we will analyse the title
of the article in terms of the noun group structure. Moreover, in this Unit we are going to
analyze the Abstract Throughout this course stress is given to the importance of
vocabulary, not just individual items but often, as in this Unit, longer stretches of language,
expressions, which are commonly found in academic writing.

Contents

Activity
1.1 The Research Article: macrostructure
1.2 Grammar focus: the noun group
1.3 The abstract

Activity 1.1 The Research Article: Macrostructure

Into how many major sections could you divide your article?
What are the titles of each section?

Read over each section again. In general terms, what is the purpose of each section and, as
such, what kind of information is included in each? Does any of the sections contain more
than one type of information?

Section Purpose // Kind of information

Abstract General overview of study

1
1.2 Grammar focus: noun groups

One outstanding feature of the titles of the articles and, indeed, of virtually all academic
texts written in English, is the use of what are called complex noun groups.
In simple terms, a noun group is a noun, either by itself or together with all the words
which help to specify it. It can act as subject, object, or complement of a clause, or as a
prepositional complement. It’s called a noun group because the word which is its head is a
noun.
Let’s take some titles of published articles. Work out the relationships between the words in
each italicised phrase. Another useful exercise would be to translate each one into Spanish
and compare the syntactic relationships of the two languages.

1. Localized Surface Plasticity During Stress Corrosion Cracking

2. Water Surface Profiles in Compound Channel with Multiple Critical Depths

3. Fuzzy Controller for Inverter Fed Induction Machines

4. Self-Organizing Feature Maps and the Travelling Salesman Problem

5. A Methodology for Solving Multi-Objective Simulation-Optimization Problems

6. The Economic Production Lot Size Model Under Volume Flexibility

Noun group structure

Determiner Pre-modifier Head Post-modifier

1. article 1. adjectives 1. noun 1. prepositional
the group
a/an
2. possessive 2. noun 2. proper noun 2. relative clause
adjective
My, your, his, her,
its, our, their
3. demonstrative 3. –ing 3. –ing clause
this, that, these,
those
4. quantifiers 4. –ed 4. –ed clause
many, a few, a lot,

2
 more, several, etc.
1. numerals 5. ’s possessive noun 5. to clause
one, two…
first, second…
6. adjective

1.3 The Abstract

Research articles in specialised publications include an abstract (sometimes called

summary) which appears between the Title and the Introduction. The abstract summarises
(normally in one paragraph) the entire text. Many academics and professionals find they
have so little free time to read everything that is published every month in their field that
they tend to look first at the titles and the abstracts in a publication. If it's interesting, they
read the whole article; if not, they go on to the next.
The abstract has thus become an important part of the article. It has its own characteristics.
More often than not, a particular publication will lay down its own “rules” with regard to
content and length. Take this example from Water Research:

Abstracts:
Abstracts have increased in importance with the growth of electronic data storage. Authors
are requested to ensure that abstracts give concise factual information about the objectives
of the work, the methods used, the results obtained and the conclusions reached.
Unsubstantiated opinions should not be included, nor should terms of technical jargon
which can only be understood by reference to the text. Where appropriate, the applicability
of the research to practice should be mentioned. A suitable length is about 250 words. For
Rapid Communications, the abstract should be limited to 100 words.

1. In terms of structure, it can follow the structure of the text itself and thus may be broken
down into the following "parts":

Antecedents (the "situation")


Problem (our knowledge is "incomplete" or imperfect")

Purpose (objectives of this study)

Method (procedures)

Results (often quantitative)

Conclusions (discusses results)

3
Of course, the abstract may present the information in a different way. This will be
especially true of articles which do not report experiments but discuss theory or describe
projects.

2. In terms of the "grammar" of the abstract, some researchers have noted the following
features:

(a) The use of a variety of verb tenses (present, past, past perfect), each corresponding to a
different part of the abstract.
(b) The use of the passive voice (methods)
(c) The preponderance of affirmative verbs and the lack of negative ones.
(d) The tendency to avoid complex sentences - few subordinate clauses.

3. In lexical and discourse terms, it is supposed that abstracts tend to:

(a) Be relatively "dense" - give a lot of information in very little space, and -
(b) Avoid repetition and exemplification.
(c) Avoid abbreviations, "jargon" and symbols.

Remember that an abstract or an article itself can break all these "rules" and still be
perfectly coherent and valid.

Is it possible to talk of “types” of abstract? Perhaps, yes. Some appear to be like “mini-
reports” of an investigation, characterised by being relatively descriptive or informative,
providing information on what was done and what the main results were. We shall call this
type informative.
Others tend to indicate in more general terms what topics are to be covered in the article.
Details are not in abundance. We shall call this type indicative.
Of course, we may find that an abstract combines both features.
Before writing an Abstract, it is worth considering what type would better suit the article
itself, the publication in which it is to appear, the type of reader for whom it is intended and
so on, and thus decide what “shape” it should take.

Analyse the abstract in your article in terms of sentence, section and key words.

Sentence Section Key words

4
 UNIT II: The Introduction

In this Unit we shall be looking at the first major section of any research article, the
Introduction.
We shall first consider the Introduction in terms of what type or types of information it
generally includes and in what order this information is generally arranged. Then, we shall
look more closely at how the first part of any Introduction is written, both from the point of
view of its content and essential grammatical features. In the next two Units (III and IV),
we shall concentrate on the rest of the Introduction in detail.

Contents

Activity
2.1 Problem-solution pattern
2.2 Moves
2.3 Text analysis
2.4 Grammar focus: verbal groups

Activity 2.1 Problem–solution texts

There are two valid ways of categorising the content of an Introduction. The first is to look
at it as a kind of mini “problem-solution” text; the second is to divide it into a series of
recognisable “moves”, each corresponding to definable discourse functions.

The problem-solution pattern of text analysis is particularly attractive to texts in

engineering, since it appears to correspond to engineering theory and practice itself, that is,
through the search for answers (solutions) to problems. It assumes that that a text is
composed of (a) a situation which requires a response and (b) a response to this situation
that is evaluated as successful or otherwise.
The situation is that part of the discourse which gives sufficient background information or
context to suggest that a problem exists. Sometimes the problem is stated as the situation
itself. The problem can therefore be thought of as that part of a situation which requires or
demands a response.
The reply or solution obviously seeks to resolve the problem in question. However,
although all solutions are responses, not all responses are solutions. A response may be
evaluated neutrally or negatively; it may be, as it were, the wrong path. In such a case, we
may return to the problem and seek another response, which may turn out to be the
solution.
If the solution is successful or otherwise, it may be evaluated in terms of its success.

5
As a rule, the situation appears at the beginning of the text, thus providing introductory,
background information which creates a context for what is to follow. However, it is
possible to find a problem and its solution before the situation. Generally, the verbs to
signal the situation are in the present simple tense.
The problem can be signalled lexically via adjectives with negative connotations (poor,
bad, inefficient, insufficient), nouns such as problem, dilemma, controversy, drawback,
disadvantage, and verbs like need, stop, prevent, avoid, not forgetting the frequency of
adversative connectors such as however, although.
Words or phrases which indicate a solution to a problem include develop, overcome, solve,
etc. The verbal form changes, thus, showing a new “functional” unit. The present perfect is
often used to signal a response, after which the text may revert to the simple present.
Lexical items often show the extent to which the response or solution was successful
(advantage, improved, more......, etc.). Sometimes we find solutions and evaluations
merged.

The problem-solution pattern of text analysis may be thus summarised as:-

1. Situation
2. Problem
3. Response + Evaluation (negative/positive)
4. Solution + Evaluation (negative/positive)

(1) What is the general topic of the article?

(2) What problems are indicated?

(3) How does the author signal these problems?

Signals
Topic
Problem
Problem
Problem
Problem

Activity 2.2 Moves

Dividing an Introduction into a series of moves assumes that the writer must include certain
types of information in the text, while other elements are optional.

Obligatory Moves:

1. Establishing the field /context

2. Outlining previous research (narrowing the field)
3. Indicating a problem/gap/need for further research

6
4. Outlining purposes

Optional Moves:

5. Announcing principal findings

6. Highlighting possible advantages/applications
7. Indicating the structure of the article

Activity 2.3 Text analysis

Read the introduction of your text. Divide it up first using the problem-solution pattern of
analysis and then according to its moves. Number the sentences in order to help you in the
analysis.

2.5 Grammar focus: verbal groups

Verbal group
Many English verbs have six forms: the Base, the –s the past, -ing participle, the –ed
participle and the to. Regular verbs have the –ed inflection for both the past and past
participle (-ed). Irregular verb forms vary.

Forms Examples Functions

base call, study, drink, put a) all the present tense, except 3rd
person singular
I/you/we/they study everyday.

b) imperative
Study.

c) subjunctive
He demanded that she study.

d) with modal verbs

He can study
-s calls, studies, drinks, puts 3rd person singular, present tense
He/she/it studies hard.
-ed1 called, studied, drank, put past tense
He called yesterday.
-ed2 called, studied, drunk, put 1. perfect tenses (to have + ed2)

b) passive voice (to be + ed2)

c) –ed2 clauses

-ing calling, studying, drinking, putting 2. progressive tenses (to be + -ing)

7
 b) –ing clauses

Infinitive to call, to study, to drink, to put infinitive clauses

(to)

8
 UNIT III: Previous Research

As we saw in the previous Unit, one very common element in practically any introduction
is the need to refer to previous research. Mention of what other researchers have done may
be made virtually anywhere in the introduction or may be included at a particular point.

Contents

Activity
3.1 Establishing the field
3.2 Previous research: how and why
3.3 Previous research: author or information prominent
3.4 Previous research: order of citations
3.5 Grammar focus: verb tenses
3.6 Discourse focus: verb tense and opinion
3.7 Vocabulary: reporting verbs

Activity 3.1 Establishing the field

In order to start the present Unit, we shall be concentrating on the first Move; that is, the
ways in which a writer establishes first what is the topic under consideration, both from a
general and a particular point of view. The text thus opens and then narrows down to
concentrate on certain aspects.

Activity 3.1 Previous research: how and why

Read the extracts and discuss the following questions:

In what ways do authors cite previous research?

In what order do they do it?
What reasons determine their choice?

A:
Division of flow past 900 branches of closed conduits (Fig.1) is encountered in flow
systems of water and wastewater purification plants (Benefield 1984) and conduit network
systems associated with heating, ventilation, and air conditioning. In the past, a large
number of experimental studies (Ito and Imai 1973; Blaisdell and Manson 1963; Gardel
1957; McNown 1954) have provided various energy loss coefficients for dividing flow in
closed circular conduits. In a very limited range of discharge ratios, Miller (1971) has
obtained the experimental loss coefficients for flow past a branching square conduit. He
asserted that the data of branching circular conduits are directly applicable to branching of
noncircular conduits. Fu et al. (1991) validated a numerical model related to three-
dimensional dividing flow in rectangular conduits using test data.

9
(taken from Dividing Rectangular Closed Conduit Flows, A.S. Ramamurthy et al., Journal
of Hydraulic Engineering, Dec.1996)

B:
…….. Gessler (1968) and Little and Meyer (1972) were among the first to systematically
study the process of bed armouring. In a series of laboratory experiments they generated
armor layers by running (over an extended period of time) sediment-free water through
straight flumes with broadly mixed (nonuniform) sediment. Based on their data, they
established relationships between the initial and final bed-material. Data obtained by
Garde, Al-Shaikh Ali, and Diette (1977) in a similar type of experiment supported
Gessler’s bed-armoring theory, which is a probabilistic approach. Garde et al. also
measured the time variation of the median-grain diameter of the bed-surface material.
Support for Garde’s approach was provided also by Lane and Carlson’s studies of armoring
in the San Luis Valley Canals in Southern Colorado.
(from Simulation of Bed Armoring in Alluvial Channels, Hong-Yuan Lee et al., Journal of
Hydraulic Engineering, Sept.1986)

Activity 3.2 Previous research: author or information prominent

There are several ways in which previous research may be mentioned. A convenient
classification would be to distinguish between those in which the author(s) are specifically
fronted (at the beginning of a sentence, as or part of the grammatical subject) and those in
which emphasis is placed on the research itself. Consider the following:

Author prominent:
1. Manson (1969) showed that ...... Researcher(s) as subject
2. Manson’s theory (1969) claims that .........
Researcher(s) part of possessive noun
phrase.
3. Manson’s (1969) theory of ............ is Researcher(s) part of possessive noun
generally accepted. phrase.
4. According to Manson (1969), the .... Researcher(s) “reported”.

Information prominent:
a.The ...........was established by Manson Researcher(s) as passive agent.
(1969).
b.. Previous research has shown that .......... Parenthetical citation.(active)
(Manson, 1969).
c.It has been shown that ...... (Manson, 1969). Parenthetical citation (passive)
d. It has been established that ............. 1-3 Superscript..
e.The ..... is probably ........... (Manson,1969). Parenthetical (with authorial comment)
f. The ............. may be ........... 1-3 Superscript (with authorial comment)

Read the following and classify the citations according to the examples given in the Tables.

1. However, recent work has shown that stress and dissolution need not be applied
simultaneously to produce brittle cracking in binary face-centered cubic (fcc) alloys.2-3

10
2. Experiments with high-sediment concentrations and laminar-flow conditions have been
performed by Ali and Georgiadis (1991), Crapper and Ali (1996), and Kusada et al. (1993,
1996).

3. As cited in ASCE (1988), Rodi (1980) listed classes of problems of interest to hydraulic
engineers that have been or can be solved by turbulence modeling, and provided guidance
for turbulence model selection appropriate for each class. While 3D models are becoming
more popular, depth-averaged models continue to be developed and used (Molls and
Chaudhry 1995).

4. The elastic stress field in an initially stress free, coated solid due to various forms of
normal loading has been analysed by, amongst others, Barovitch et al.1

5. Howard Gardner’s (1983) theory of multiple intelligences has great potential for helping
revolutionize our concept of human capabilities. Gardner’s basic premise is that
intelligence is not a single construct: individuals have at least seven distinct intelligences
that can be developed over a lifetime.

Activity 3.3 Previous research: order of citations

In terms or order, there seem to be 3 ways of presenting the findings of previous research.

1. Strictly chronological (earliest to latest)

2. Distant to close (from general findings in the area to particular research related directly
to the study)

3. Grouped by approaches (different research paradigms or interpretations)

Read the following Introductions and decide if they follow pattern 1, 2 or 3.

1.
INTRODUCTION
A vast amount of effort has already been dedicated to the study of supervised learning
algorithms, such as the Boltzman Machine (Hinton, Sejnowski, & Ackley, 1984), and the
Back-Propagation algorithm (Le Cun, 1987; Rumelhart, Hinton, & Williams, 1986). Less
attention has as yet been devoted to the class of algorithms which do not require explicit
tutoring by input-output correlations and which spontaneously self-organize upon
presentation of input patterns. This whole area of research is highly promising and directly
relates to the difficult problem of understanding the internal representation of information
in the brain, in particular, how structures and internal organization can emerge from the
collective behaviour of interconnected neuron-like elements (Linsker, 1986; von der
Malsberb, 1985).
Self-organizing feature maps (Kohonen, 1984) offer such a paradigm, by setting up some
general functional principles that could be responsible for the self-organization of widely
different kinds of information. Kohonen shows how input signals of arbitrary
dimensionality can be mapped, or adaptively projected, onto a structured set of processing

11
units, in such a way that topological relations of the input patterns and of the representation
patterns are kept similar. Algorithmically, his method can be seen as an extension of
competitive learning (Rumelhart & Zipser, 1985), in which clusters are given a topology
organizing them into 2-D structures (such as trees, rectangular or hexagonal lattices). When
for a given pattern, the best matching unit is found, not only its own weights but the
weights of its topological neighbours are modified to increase the strength of the match.
Applications to various cognitive tasks have been demonstrated by Kohonen. We intend to
show the potential of this approach for optimization problems as well, by applying it to the
well-known Travelling Salesman Problem (TSP).
(Self-Organizing Feature Maps and the Travelling Salesman Problem, Bernard Angéniol et
al., Neural Networks, Vol.1, 1988)

2. (extract-first paragraph only from Model-free forecasting for nonlinear time series,
Berlin Wu, Computational Statistics &Data Analysis 19, Elsevier Science, Holland, 1995)

To exhibit the empirical data in an appropriate model in the application of nonlinear time
series, researchers in this field have been concerned with the nonlinear system such as
bilinear models (Granger and Anderson, 1978; Subba Rao and Gabr, 1984), threshold
autoregressive models (Tong and Lim, 1980), state-dependent models (Priestley, 1980),
autoregressive models with conditional heteroscedasticity (Engle, 1982), autoregressive
models with discrete state space (McKenzie, 1985) and doubly stochastic models
(Tjostheim, 1986). Their works draw light on closer understanding of the nonlinear
phenomena underlying the data (c.f. De Gooijer and Kumar, 1992). However, these works
are complicated issues, solutions to some of which have eluded us even in the case of linear
models, and promise to be far more complicated when we enter the realm of nonlinear
dynamic models, see Granger (1991) and Pesaran and Potter (1992).
......... (The goal of this paper is to propose a neurocomputing technique to perform the
robust forecasting for some nonlinear time series. ......)

Activity 3.4 Grammar focus: verb tenses

Read the text and decide which of the verb forms in parenthesis is “correct”.

Text Generation:

Computer text generation (is/was/has been) the process of constructing text (phrases,
sentences, paragraphs) in a natural language - in a sense, it (is/was/has been) the opposite
of natural language understanding by machine. Although this problem (is/was/has been)
investigated for 15 years, few coherent principles (emerge/emerged/have emerged), and the
approaches (vary/varied/have varied) widely. Attempts at generating text (are made/were
made/have been made) with two research goals: (a) generating random sentences to test a
grammar or grammatical theory and (b) converting information from an internal
representation into a natural language.

12
Random Generation
This approach, the random generation of text constrained by the rules of a test grammar,
(is/was/has been) of limited interest to workers in Artificial Intelligence, since it
(is/was/has been) oriented more toward theoretical linguistics than toward functional
natural-language-processing systems. The objective of implementing a generation system
of this sort (is/was/has been) to test the descriptive adequacy of the test grammar, as
illustrated by the following two systems.
Victor Yngve (1962) (is/was/has been) one of the first researchers to attempt English text
generation; the work (is seen/was seen/has been seen) as preliminary to a full program for
machine translation. Yngve (uses/used/has used) a generative context-free grammar and a
random-number generator to produce “grammatical” sentences: The system randomly
(selects/selected/has selected) one production from among those that (are applicable/were
applicable/have been applicable) at each point in the generation process, starting with those
productions that (“produce”/ “produced”/ “have produced”) SENTENCE and finally
randomly selecting words to fill in the NOUN, VERB, and other like positions. This
(is/was/has been) an example of the text produced by the system:

The water under the wheels in oiled whistles and its polished shiny big and big trains is
black.

(taken from The Handbook of Artificial Intelligence, Vol.1., p.273, Eds.Avron Barr & E.A.
Feigenbaum, HeurisTech Press, Stanford, California, 1981)

Activity 3.5 Grammar focus: verb tense and opinion

Without doubt, the three most common tenses used in the writing of introductions are the
Present, the Simple Past and the Present Perfect. These terms may be misleading since their
use is not necessarily restricted to time but also to how we relate events together and “see”
them in terms of “now”. In other words, if we consider:

(a) Kolkovski pointed out that the main problem lies in ........

(b) Kolkovski has pointed out that the main problem lies in ........

(c) Kolkovski points out that the main problem lies in ..........

the most typical interpretations of the three statements would not be in terms of increasing
“present-ness” but in terms of increasing generality and/or relevance

1. Generalisations (information prominent- researcher not mentioned): tend to be in the

Present tense.
2. References to specific experiments (researcher mentioned/reference to a single study):
tend to be in the Past tense.
3. References to areas of inquiry (researchers often mentioned/reference to several studies):
tend to be in the Present Perfect tense.

13
To put this in another way, we could say that:

1. The Present tense and also the Present Perfect tense are used to claim generality about
past research. The Past tense is used to claim non-generality about past research.
2. The Past tense is used to refer to the quantitative results of past research. Sometimes it is
used to show that the research mentioned is non-supportive of some or all the aspects of the
work described in the article. The Present tense is used when it refers to research which is
supportive.
3. The Present Perfect tense is used to indicate the continued discussion of some aspects of
the research.

Activity 3.6 Vocabulary: reporting verbs

As we have seen above, the choice of one tense over another may reflect something of the
author’s attitude towards his/her subject.
Another way in which attitude is reflected is through the verb chosen to report previous
research.
Verbs can either be neutral or carry some kind of authorial commitment or attitude towards
other research. The author may thus signal whether claims are to be taken as substantiated
or not.

A:
Read the following series of short extracts which illustrate the use of some common
reporting verbs. Decide in each case whether the verb is neutral or suggests some kind of
authorial comment.

show
Coats and Smith 6 showed that Dean’s model can be reformulated to include the effects of
mass transfer.

develop
Ashley et al. (1980) developed a procedure for testing whether the difference between the
mean square errors obtained from any two models is significant.

examine
Wang examined the effects of the different parameters in antenna measurements and how
they affect FF pattern reconstruction by mode-sensitive algorithms. (3)

indicate
Previous work on the chemistry of metals has indicated that lake sediments act as a sink for
these substances (Oschwald, 1972).

suggest
Treitel (1981) suggested that the most significant reason for this poor reputation is the
excessive sensitivity of homomorphic filtering to noise.

use

14
Researchers in neurocomputing use feedforward networks, e.g. Lapedes and Farber (1987);
Kosoko (1991); to predict future values of time series only by extracting knowledge from
the past.

propose
Phadke et al. (2) propose the Discrete Fourier Transform of the voltage samples to be used
recursively for calculation of a stationary phasor, and positive sequence phasor rotation to
be used for measurement of the frequency.

argue
As a filtering or smoothing process, Gallagher and Wise (1981) argued that median
filtering, when applied to a time series consisting of signal plus noise, produces a root
signal which is one or the other.

report/investigate/study/present
Several numerical studies of scattering by a crack in an infinite or semi-infinite medium
have been reported. By using the boundary integral equation, the scattering of incident
surface waves and incident body waves by a surface-breaking crack in an elastic half-space
was investigated by Mendelsohn et al. The case of a subsurface crack was studied by
Achenbach and Brind (1981) for scattering of Rayleigh waves. An approximate solution
that is valid at low frequencies was presented by Datta (1979) for SH wave diffraction by a
canted surface-breaking planar crack.

Can you think of another way to classify reporting verbs?

Activity 3.7
Take the article you have decided to analyze. This time continue by determining how the
author refers to relevant previous research.

15
 UNIT IV: from PROBLEM to PURPOSE

Almost invariably an author will mention at some stage of an Introduction that some kind
of “problem” or “gap” exists. You may remember in Unit II that one way of analysing an
Introduction is through the “problem-solution” pattern and that often the “situation”
outlined at the beginning of the text may contain or may be a “problem”. However, in this
Unit we are more concerned with the way in which a writer can highlight “problems” later
on in the Introduction. This could either be done in a relatively “linear” fashion, i.e. after
commenting upon previous research efforts, the writer can then point to their limitations,
inherent problems etc. Alternatively, given a more complex situation perhaps, the writer
could adopt a “cyclic” approach, mentioning a series of attempts to tackle or “solve” the
problem, each followed by comments as to their drawbacks.
Pointing out limitations, problems and so on leads to a part of the Introduction which is
obligatory: announcing what the aims of the research are. In a metaphorical sense, it is as if
the writer is creating some kind of personal territory for him or herself by saying that - to
date - this problem has not been solved or that there is still work to be done. This territory
will presumably be then occupied by the study to be described.

Contents

Activity
4.1 Gaps, problems and traditions
4.2 Text analysis: classifying “problems”
4.3 Discourse focus: adversative connectors
4.4 Discourse and vocabulary: expressing purpose

Activity 4.1 Gaps, problems and traditions

It is an over-simplification to talk only in terms of problems. The writer may want to point
to any of the following:

1. Previous research is inadequate because an important aspect of the research has been
totally or partially ignored. (gap)
2. There is an unresolved conflict among the authors of previous research
(theoretical/methodological disagreement). (problem)
3. Previous research suggests an extension or new research question emerging from it.
(continue tradition)

A:
Read the Introduction of your article and:

(a) decide if and where the writer is indicating a gap, a problem or a continuation of a
research tradition.

16
(b) underline which words or expressions the writer uses to indicate any of the above.

(c) indicate what the purpose of the study is.

Activity 4.2 Text analysis: classifying “problems”

Read the following set of extracts from Introductions. Classify each one

1. In the calculation of lightning voltages on transmission lines, the accurate representation

of transmission tower is a difficult problem and has been the subject of much discussion.
Several attempts (2-4) have been made to experimentally measure the tower surge
impedance on full-scale towers. However, the resultant values for surge impedance were
found to be inconsistent both in magnitude and in time variation.

Gap / Problem / Continuation

Linguistic signals: _______________________________________________________
______________________________________________________________________

2. Despite some field use of this type of diverting agent as an acid preflush and despite its
potential use for long-term water control, relevant published data on the subject are
scarce.13

Gap / Problem / Continuation

Linguistic signals:________________________________________________________
______________________________________________________________________

3. Although the kinetics of bacterial growth and substrate removal are known to play an
important role in the design of activated sludge waste treatment processes, relatively little
attention has been given to investigating the effects of changes in the mixed populations on
the kinetic parameters.

Gap / Problem / Continuation

Linguistic signals:________________________________________________________
______________________________________________________________________

4. Although there is a wealth of rain attenuation data in the literature (1-3), recent
investigation (4-6) all point to the inaccuracy of these prediction models for use outside
Europe and North America.

Gap / Problem / Continuation

Linguistic signals:________________________________________________________
______________________________________________________________________

5.The need to alter the structure or the chemistry of the surfaces of the metallic implants in
order to improve their functions or to reduce their bioreactivity has not been generally
recognized. Few studies have been conducted on potential carcinogenic activities or

17
immunological responses to the major alloying elements of medical devices(2). For surgical
applications, 316L stainless steel has been replaced by emerging materials which are more
electrochemically stable but which, as in the case of titanium and its alloys, are prone to
undesirable wear.

Gap / Problem / Continuation

Linguistic signals:________________________________________________________
______________________________________________________________________

6. Up to now, there have been very few studies on the effect of floor slabs as diaphragms on
the building response; there is no design procedure which considers the diaphragm action
of floor slabs. The lack of information regarding the in-plane characteristics of floor slabs is
primarily responsible for this situation.

Gap / Problem / Continuation

Linguistic signals:________________________________________________________
______________________________________________________________________

7. Most of the techniques used involve some large set of equations to be numerically
solved. In all data so far published in the literature, nothing or very little has been said
about the form of the field inside the structure.

Gap / Problem / Continuation

Linguistic signals:________________________________________________________
______________________________________________________________________

8. In the past the bulk of the research effort has been devoted to crack propagation.
Considerably less attention has been given to the initiation process.

Gap / Problem / Continuation

Linguistic signals:________________________________________________________
______________________________________________________________________

9. The issue of turbulence model selection is well documented in the literature, yet the
number of published applications of turbulence-based models to real-engineering complex
geometries remains small.

Gap / Problem / Continuation

Linguistic signals:_______________________________________________________
______________________________________________________________________

Activity 4.3 Discourse focus: adversative connectors

18
Typical markers to criticize previous research are the adversative connector (however,
nevertheless, yet, unfortunately, but); lexical items - especially verbs (suffer, is limited to);
adjectival phrases (time consuming, expensive) and verbal negation (cannot treat).

Adversative Connectors:

The function of adversative connectors is to point out contrast, difference, opposition. We

are talking about words and phrases such as:

on the contrary, on the one hand...on the other hand, by contrast, alternatively, instead, then,
however, nevertheless, though, conversely, in comparison, rather, but, whereas, in spite of,
despite (the fact that), although, even though, though, yet

If we take the following sentence

(Adversative connector:) Although petroleum consists basically of compounds of only two

elements, carbon and hydrogen, these form a large variety of complex molecular structures.

we can see that it consists of two propositions:

(a) petroleum consists basically of only two elements, carbon and hydrogen
(b) carbon and hydrogen form a large variety of complex molecular structures.

The two can be joined in a number of ways:-

1. ALTHOUGH petroleum consists basically of only two elements, carbon and hydrogen,
these form a large variety of complex molecular structures.
2. Petroleum consists basically of only two elements, carbon and hydrogen; HOWEVER,
these form a large variety of complex molecular structures.
3. Petroleum consists basically of only two elements, carbon and hydrogen; these,
HOWEVER, form a large variety of complex molecular structures.

As you can see, although and however take up different positions in the sentence. They may
be similar in semantic terms, but their syntactic or grammatical use differs. A useful
classification of adversative connectors would thus be:

Sentence connectors Subordinators Co-ordinators Others

However although but despite (+ noun)
Nevertheless even though yet in spite of (+ noun)
Nonetheless though compared with
in contrast while unlike
in(by) comparison whereas
on the other hand
on the contrary
Still

19
Notes:
(1) Sentence Connectors:
There are two types: phrases, such as in contrast and on the other hand, and conjunctive
adverbs, such as however and nevertheless.
Both usually appear at the beginning of sentences and are separated from the rest of the
sentence by commas.

In the United States, the president may not belong to the same political party that holds a
majority in Congress. In contrast, the British prime minister is the head of the party that
has most seats in Parliament.
(alternative: ....The British prime minister, in contrast, is the ..... (after the subject)

I would like to study law. However, my family want me to be an engineer.

(alternatives:- My family, however, want me to be an engineer. (after the subject); My
family want me to be an engineer, however. (at the end)

We can also join two clauses together with a conjunctive adverb using a semi-colon and a
comma.

I would like to study law; nevertheless, my family want me to be an engineer.

(2) Co-ordinators and Subordinators

Co-ordinating conjunctions are used with a comma to join two independent clauses to form
a compound sentence. The difference between but and yet is that the latter is preferred
when the second clause is an unexpected or surprising contrast to the first clause.

I want to study art, but my parents want me to become an engineer.

Winston Churchill needed extra English classes at school, yet he went on to win the Nobel
prize for Literature.

Subordinating conjunctions introduce a dependent (or subordinate) clause. This clause can
either be placed first:

Although the film was very popular with the general public, film critics thought it was
terrible.

or after the main clause:

Critics thought the film was terrible although it was very popular with the general public.

(Notice the use of the comma in the first, but not in the second.)

(3) Others
What is most important here is to remember to follow these expressions with a noun:

Despite the film’s popularity with the general public, critics thought it was terrible.
Unlike a human worker, a robot can work 24 hours a day.

20
As a gerund is, functionally, a type of noun, it too can be used:

In spite of having a memory, a robot has to be programmed by a human.

It is important to remember that with these “others”, no conjugated verb (i.e. one with
person and time) can appear in the same clause as them.

Choose the article of interest to you. Analyse those parts of the Introduction that point to a
“problem” and decide whether it refers to a gap, a problem or to the continuation of a
research tradition. What linguistic resources are used - adversative connectors, verbal
negation, lexical items?

Activity 4.4 Discourse and vocabulary: expressing purpose

In all articles, stating what the specific aims of the study are is an obligatory element. There
are two predominating forms:

1. The author(s) indicate their main purpose or purposes.

2. The author(s) describe what they consider to be the main features of their research.

Notes:
(a)One predominant feature of both is the use of references to the text itself through words
like this paper, the present ........, this article.

(b)A second characteristic involves the choice between a standard or collapsed form when
introducing purpose. Compare:-

The purpose of this paper is to describe .....

This paper describes .....

(c)A third point of note concerns the use of tense. In cases where the reference is the genre
(paper, report, review, note, article) tense is restricted to the present. In cases where the
reference is the type of inquiry (investigation, study, experiment), you can choose between
present and past.

The purpose of this investigation is to ........

The purpose of this investigation was to ......

A:
Look at the following, taken at random from articles published relatively recently (Nov.-
Dec.1996) and (a) distinguish between standard and collapsed forms, (b) note the verb
tenses used.

1. In the present paper, the ............................................ is examined.

2. The main aims of the project were to .................
3. The purpose of this article is to ...............................

21
4. The purposes of this paper are to ........................
5. In this study, the ........................................ was assessed.
6. The main objectives of this research were to ............................
7. This paper reports on laboratory experiments that were conducted to address .............
(the)
8. This paper presents results of ............................
9. The present study deals with the .................................
10. This paper describes theoretical and experimental investigations of the surge response
of a transmission line tower, undertaken to clarify its representation in the calculation of the
lightning performance of a transmission line.

11. In this study, the effects of dilution rate in continuous mixed cultures of sewage origin
on the growth behavior of the populations enriched in the system were investigated.

12. The present study was begun in order to address specific field problems and the general
lack of knowledge pertaining to this type of diverting agent. The goal was to achieve a
better mechanistic understanding of the overall injection, reaction, and subsequent
diversion processes that occur in both water- and oil-bearing strata during treatment.

B:
Choosing appropriate “openings” from the above examples or from your own resources,
write sentences to complete the following.

1. (Ambiguity in multicriteria quality decisions)

....analyze quality from a decision-making perspective.
_________________________________________________________________________
___________________________________________________________________

2. (Impact of Landscape Feature and Feature Placement on Agricultural Non-

Point-Source Pollution Control)
.....assess the cumulative impact of width and placement of vegetated buffer strips on
water quality in the Bluegrass watershed in southern Iowa.
_________________________________________________________________________
___________________________________________________________________

3. (Sensitivity based expert system for voltage control in power systems)

.....develop an expert system for voltage control taking advantage of the above methods
and with added features to overcome the deficiencies of the earlier methods.
_________________________________________________________________________
___________________________________________________________________

4. (The Economic Production Lot Size Model Under Volume Flexibility)

......formulate and solve an EPL model that takes into account the following:

22
 1. Production rate is a decision variable.
2. Unit production cost is a function of production rate.
_________________________________________________________________________
___________________________________________________________________

5. (Fuzzy Controler for Inverter Fed Induction Machines)

......a fuzzy controler for stator direct self control of an induction machine.
_________________________________________________________________________
___________________________________________________________________

6. (Experimental Study of Beam-Supported Slabs Under In-Plane Loading)

......work done at Lehigh University as part of a comprehensive research project on the
contribution of floor slabs to the earthquake resistance of building structures.
_________________________________________________________________________
___________________________________________________________________

Again take the article you are working with and note down how the authors state their
research aims.
Also, pay particular attention as to how the Introduction continues after the aims have been
highlighted, if it does.
Remember in Unit II how optional moves included

Optional Moves:

5. Announcing principal findings (results)

6. Highlighting possible advantages/applications (or even limitations?)
7. Indicating the structure of the article

UNIT V: METHODS AND MATERIALS

In this Unit, we shall be concentrating on that part of the research article which at least
appears to be the easiest to write, namely, the section often called Experimental, Procedure
or, in our case, simply, Methods and Materials.
It is simpler perhaps because it is more descriptive; however, as we shall see in some cases,
the information contained in this section is often very dense and requires considerable
content knowledge to fully understand it.

Contents

Activity
5.1 Methodology section: features
5.2 Grammar focus: the passive voice
5.3. Discourse focus: specialized knowledge

23
 5.4 Methodology section: content
5.5 Materials

Activity 5.1 Methodology section: features

The methodological or experimental section of many articles contains recognisable
linguistic features. Read through the following two examples and try to work out what two
of these features are.

A:
STRAIN HARDENING OF ALUMINIUM AT HIGH STRAINS

A. Korbel, F. Dobrzanski, & M.Richert

EXPERIMENTAL

(1) The investigations were carried out using commercially pure aluminium (99.5% Al). (2)
The aluminium ingots were reduced by hot rolling to the thickness of 12 mm.(3) The
temperature at the beginning of the hot rolling was 783 K (510 0C). (4) The differences in
the value of the preliminary strain, and thus the different degree of strain hardening was
obtained through cold rolling. (5) Table 1 gives the scheme of the applied reductions and
the corresponding values of the true (logarithmic) strain. (6) The hot and cold rolling were
carried out under industrial conditions.
(7) The properties of the material after hot rolling and the effect of cold deformation were
determined on the basis of tensile tests and structural observations conducted by using both
optical and electron microscopy.
(8) The samples for the tensile tests were cut out parallel to the rolling direction. (9) The
tensile tests were carried out at room temperature and at a strain rate of 10--4s--1.
(10) The number of samples cut out from a rolled band after each reduction and examined
in the tensile test was 60. (11) This made possible a statistical study of the experimental
results. (12) The following parameters were used for the description of a given feature X:
.............

Acta metall., Vol.31, No.2, 1983.

B:
MATERIALS AND METHODS

(1)A field and laboratory study was conducted on a Hord silt loam, the former at the
Management Systems Evaluation Area (MSEA), Shelton, NE, and the latter on intact soil
cores, obtained from the same site, but incubated in the laboratory. (2)Some characteristics
of the soil are given in Table 1.
Field Study

24
(3)Six PVC cylinders (28.7 cm i.d. by 1.8 m) were individually pushed into the soil to a
depth of 1.2 m using a metal collar placed on top of the cylinder and the bucket of a large
front-end loader, which provided about 10,000 kg of down loading from the loader bucket
placed on top of the collar. (4)A trench (2.1 by 2.1 by 4.2 m) was dug next to the columns
so that one side of the columns was exposed and timber shoring erected so that the columns
were held in positions for the duration of the study. (5)Gas probes were inserted through
holes, drilled in the walls of the cylinders, into the center of the soil column, at depths of
30, 60, 90, and 120 cm from the soil surface. (6)Distilled water (11.1 L) containing 30 mg
L-1 NO3-N (76.6 atom% 15N) was added to the top of all columns while three columns also
received an addition of 10 mL of ethanol with the NO3 solution. (7)These additions
corresponded to application rates of 51.8 kg NO 3-N ha-1 and 644 kg C ha-1, respectively.
(8)The ethanol requirement was calculated according to McCarty et al. (1969) using the
concentration of NO3 in the soil profile to a depth of 120 cm before the application of the
treatments (mean value of 31.7 mg N kg -1 soil) and the concentration of NO3 in the added
solution. (9)The solutions were allowed to percolate through the columns over 4 d. (10)The
headspace above the soil in the column was sampled every 6 h by capping the column for 1
h and taking gas samples after 0, 30 and 60 min. (11)Gas samples were also obtained from
the gas probes within the soil columns after 6 h. (12)At the end of 4 d, three soil cores were
taken from inside each of the soil columns to a depth of 1.2 m and bulked across depths in
30 cm increments.
(13)The gas samples were analyzed for N 2O and CO2 by gas chromatography (Weier et al.,
1993) and for 15N-(N2O + N2) by isotope mass spectrometry (Mosier et al., 1986).
(14)Fluxes of 15N-(N2O + N2) from the soil surface within the columns were calculated as
described by Hutchinson and Mosier (1981). (15)The results were corrected for the
solubility of N2O and CO2 in soil water (Moraghan and Buresh, 1977)
(Water Quality: Potential for Bioremediation of High Nitrate Irrigation Water Via
Dentrification, KL Weier et al., Journal of Environmental Quality, 23:105-110 (1994).)

Activity 5.2 Grammar focus: the passive voice

Perhaps the two most striking linguistic aspects of the above texts are that practically all the
sentences begin with nominal groups (i.e. a noun or noun group) and that the verbal form
used in the passive voice in the past tense.

(7) The properties of the material after hot rolling and the effect of cold deformation were
determined .....
(3)Six PVC cylinders (28.7 cm i.d. by 1.8 m) were individually pushed into the soil ...

Before we go any further, let us first look at the passive voice.

In very general terms, the passive voice is used if the agent (the subject) is considered less
important than the action or if the agent is very general (we, they, the people, the police) or
simply unknown (someone, something).

In describing processes/experiments the passive voice is also frequent, the agent almost
invariably being the researcher(s).

In terms of form, the following more difficult examples may be useful:

25
 Active Passive
Infinitive to choose to be chosen
Perfect Infinitive to have chosen to have been chosen
Participle and Gerund choosing being chosen
Perfect Participle and having chosen having been chosen
Gerund

As for more information on when to use it, the following notes should provide the
necessary information:

1. When the active form would mean the use of an indefinite or vague pronoun or noun as
subject.

(a) I’ve been robbed! (Someone has robbed me!)

(b)The building had to be demolished. (They had to demolish the building.)
(c) It is assumed that the Government will do something about the situation.
(People assume that the Government will do something about the
situation.)
2. To avoid a change of subject in the middle of a sentence.

The President stepped off the plane and was immediately surrounded by reporters.
(The President stepped off the plane and reporters immediately surrounded him.)

3. To make a statement sound impersonal, either out of modesty or to avoid personal

responsibility.

(a) The new working methods that are to be introduced may result in some staff cuts. (This
is bad news for the workers, but it’s not really our fault!)
Compare this with:-
(b) The new working methods we are introducing will result in higher salaries for all
workers. (Don’t forget to take off your hats when you see us!)

4. The passive is therefore not just an equivalent alternative to the active. While both may
be syntactically possible, we tend to choose the passive for one of the reasons above or if
we are interested in what happened to “x” rather than what “y” did. Thus:-

(a) The murderer was arrested 2 days later. would be preferred to The police arrested the
murderer 2 days later. if we are more interested in what happened to the murderer, telling
the story, as it were, from his (or her?) point of view.

Similarly,

(b) Several trees were struck by lightning in last night’s storm. would probably be preferred
to Lightning struck several trees in last night’s storm. since we are likely to be more
concerned with what happened to the trees than in what the lightning did, although in this
case the agent (by the ......) is included.

26
5. Remember, though, that sometimes either the active or the passive may be used, and the
choice is simply one of “focus of interest”. Take:-

(a) Manchester United beat Liverpool by 3 goals to 1.

(b) Liverpool was beaten by Manchester United by 3 goals to 1.

A:
Read the texts in 5.1 again. Underline all the examples of the passive voice.

Activity 5.3 Discourse focus: specialized knowledge

What is also noticeable about the methods section is that in some areas, the description
seems to depend very heavily upon the knowledge the reader brings to the text. The
sentences are like stones across a river, stones which only the specialised reader can jump
from one to the next to reach the other side.

For example, in the first text in 5.1 the naive reader might ask, what is hot rolling? what is
cold rolling? what are the industrial conditions under which these rollings were made?

The following is by no means an extreme example of this. Read through it and underline
those parts of it you consider to be only comprehensible to the “expert” reader.

Experimental Analysis

Total organic carbon (TOC) was analyzed on a Shimadzu TOC analyzer Model 500
(Shimadzu Co., Japan). Total TOC of the effluent was analyzed after the suspended
bacterial flocs in the sample were completely disrupted in a Tomy Ultrasonic Disrupter
Model UD-201 (Tomy Seiko Co. Ltd, japan). Soluble TOC of the effluent was analyzed
after filtering it through a 1 m membrane filter (Advantec, Japan). The difference between
total and soluble TOC gave the effluent suspended solids TOC (SS-TOC) concentration.
Methanol and gas composition were measured by a gaschromatograph using FID and TCD
respectively as elaborated previously (Bhatti et al., 1993b). Low molecular weight VFA´s
were analyzed gaschromatographically using a Hitachi 263-50 GC (Hitachi Co., Japan)
equipped with 2 m x 3 mm (i/d) glass column packed with Unisole F-200 (30/60 mesh).
The column temperature was 1400C and that of injection and FID port was 180 0C. Nitrogen
was used as the carrier gas flowing at a rate of 37 ml/min. Hydrogen partial pressure was
determined by detecting the amount of H2 in the gas phase. For this purpose, samples taken
in an air-lock syringe were analyzed on a Hitachi GC model 063 (Hitachi Co., Japan) with
TCD equipped with 3 m x 3 mm (i/d) stainless-steel column packed with a molecular seive
13X (80/100 mesh). The carrier gas was nitrogen at a flow rate of 32 ml/min. Column
temperature was 420C while injector and detector ports were at room temperature. All other
analyses were carried out according to Standard methods (APHA 1985).

(Feasibility of Methanolic Waste Treatment in USAB Reactors, Z.I. Bhatti et al. , Water
Research, 1996.)

27
Activity 5.4 Methodology section: content
Turning our attention now to the content of a typical experimental section, many of the
following aspects will no doubt be included:

1. Overview
2. Sample/Population/Location
3. Restrictions (Variables)
4. Sampling Technique
5. Materials*
6. Procedure*
7. Statistical Treatment

Although the two marked sections (*) are alone in being virtually obligatory, any or most of
the others will often be found, especially in studies that are based on laboratory or field
experiments.

Read the second text in 5.1 again (the one on Water Quality on pages 44-45), and then the
text which follows the Table and identify what aspects are included, noting down which
sentences correspond to which features.

Aspect Text 1: Water Quality Text 2: Frequency ....

Overview
Sample/Pop./Location
Sampling Technique
Materials
Procedure
Statistical Treatment

Text 2:
The three methods discussed in sections 3-5 have been tested by computer simulation
Several scenarios have been tried, and their performance documented. Whenever the term
true frequency is used, it refers to the quantity (2) obtained from the signal, which is
represented in the form (1). Simulated transients are disturbances of the waveform
represented by the equation (1), with the following modulations and disturbances added:

∙ The tracking abilities need to be tested under transient conditions. A 1 Hz swing was
modulated on the nominal frequency, with a maximum value of 1 rad/sec.

∙ Various amounts of measurement noise, modelled according to (2), were added to the
signal. Typical values of standard deviation are of the order of 1%.

28
∙ Quantization noise was added to the samples. It corresponds to the 12-bit A/D converters
used in the phasor measurement system. the sampling rate was 1440 Hz (24 samples per
cycle), which allows for 12 harmonics to be present in the signal without aliasing effects.

∙ A subsynchronous oscillation of 6 Hz was modulated on top of the transient swing of 1

Hz. Even though this type of oscillation is not very common, it would test the tracking
performance of the proposed methods in that important frequency region.

∙ Various amounts of harmonics were used in the signal. Three scenarios are presented in
the paper: 5% 3rd. harmonic, 5% distortion from the harmonics (3,5,7,9,11), and 25%
distortion from the same group of harmonics.
(Frequency Tracking in Power Networks in the Presence of Harmonics, Miroslav M.
Begovic et al., IEEE transactions on Power Delivery, Vol. 8, No. 2, April, 1993)

Activity 5.5 Materials

If a fairly detailed description is required of the materials used, some of the following
features may be present.

Materials:
laboratory equipment
field equipment
human/animal subjects
natural substances
fabricated materials
surveys/questionnaires
computer models
mathematical models

Which parts of the following two texts have recognisable features corresponding to some of
the above?

Text 1:
Application of Fuzzy Control in Chemical Distillation Processes
(Gerhard Klett, IEEE, 1993)

Abstract: Fuzzy control, combined with common control techniques, offers new
possibilities to implement descriptive models for complex control tasks in a brief and
distinct manner. This paper describes the application of fuzzy control in a non-continuous
chemical distillation process.

I Introduction:
Distillation is used widely in the chemical industry to separate components with different
boiling temperatures of a liquid. The distillation apparatus generally consists of a
distillation flask, a packed distillation column, a condenser and valves to adjust the reflux
ratio and for the collection of the separate components. The distillation flask is placed in a

29
heating bath with adjustable temperature. Sensors for pressure and temperature are attached
to the apparatus. Fig. 1 is a schematic of the distillation apparatus.

The goal of distillation is the separation of the mixture into its components with high purity.
A manual control process is performed in the following way: depending on the temperature
and pressure gradients at several locations inside the apparatus. The temperature of the bath
TB, the pressure p and the reflux ratio are varied inside the apparatus. Additionally, the
human expert bases his control strategy on some non-quantifiable phenomena such as the
shape of the drops at the column head or the occurrence of “streaks” in the column.

II Fuzzy Logic Based Model:

For the configuration of a fuzzy control we must first extract linguistic variables (LV), their
values and if/then rules from the description a human expert. For controls, which form a
closed loop within the process, it is a basic presumption to have linguistic variables with
related technical values. Each value of a linguistic variable represents a fuzzy set. The
degree of membership of a technical value x to a fuzzy set is determined by a so-called
membership function µ(x) ε[0,1] (see [1] and [2]for more details about fuzzy set
theory. Fig. 2 gives an overview of a fuzzy inference scheme.

For the design of the control strategy, two things are important: the separated components
of the mixture should be as pure as possible and the stress of the apparatus, expressed as
pressure in the apparatus, should be constant. From Physics it is known that a constant
temperature at the column head T3 (see Fig.1) is equivalent to a constant pressure during the
separation of a component. An increase in this temperature identifies the separation of the
next component. Fig.3 shows the graph of T3 over time for the separation of two
components.

After the knowledge acquisition, by interviewing the experts, we could isolate eight
linguistic variables and their values: five for input and three for input (Table 1).
We use linear membership functions in our model, giving the technical values linear
weights [3]. Fig.4 shows the membership functions of bath temperature and pressure.
As mentioned above, the expert uses two non-quantifiable values, streaks and dropshape,
for his control strategies. A mathematical model was developed to substitute measurable
signals for these subjective phenomena. Fig.5 shows the structure of the fuzzy control with
input and output variables.

(the remainder of the text is divided into the following sections:)

III Hard and Software Development Environment
IV On-Line Connection to the Process
V Software Structure of the Fuzzy Distillation
VI Results
VII Conclusions

Text 2:
The Equipment

30
The system is basically constituted by two main units: the remote station and the central
system. The remote station includes the sensor, the signal conditioner, controller, radio and
batteries. the central system is formed by the radio and a personal computer.

The remote station:

As stated above, because dike J is situated in an area of very difficult access, it was
necessary to choose a capacitive type gauge requiring little maintenance.
It was decided to build the gauge based on two commercial aluminium concentric tubes of
3.2 m. long. The diameter of the internal tube is 1.25 cm. and the external one 6.35 cm.
with a thickness of 5 mm. This ensures that the gauge is strong enough to resist any impact
by floating bodies. On the external tube, 25 cm. x 1 cm. slots were made along its length so
as to guarantee easy access of the fluid and, consequently, an appropriate dynamic
response. To reduce the parasitic effects of fluid conductivity, both of the tubes were
anodized and electrostatically painted. To evaluate the dynamic response of the gauge, a 1
m. model was tested in a 2-D laboratory wave tank with periodic waves. Its output was
compared with a commercially available transducer and no important differences were
observed.
....
(taken from Remote Water Wave Measurements Using A New Capacitive Transducer, by H.
Navarro et al., III International Congress on Energy, Environment and Technological
Innovation, Proceedings, Vol.3, ps.493-498, 1995)

Activity 5.6 Study check

In the text of your interest, analyse the experimental section from a grammatical and/or
analytic point of view.

UNIT VI: RESULTS

In most research, once the data have been analysed and the results obtained, the last phase
of the research process is to summarise the results and interpret them. It is at this stage that
the researcher puts the results together and then attempts to understand their meaning
within the general context of the research problem and topic. Questions may be asked (or
even stated) such as:

What were the major findings of the research?

What do these results mean?
What can be learnt from them?
What are their implications?
How can they contribute to existing knowledge in the research area?
What recommendations can they lead to?

Research articles sometimes differentiate between reporting and summarising results on the
one hand, and interpreting them on the other. In the former, often simply called the Results
section, what we tend to see is a report of the results obtained from the analysis of the data,
usually through the use of tables, graphs, charts, lists, and so on, together with a clear

31
synthesis of the findings. As for interpretation, the researcher goes beyond the results
themselves towards conclusions, implications and recommendations based on the results. In
many research papers, this section is referred to as the Discussion.
However, some articles do mix the two sections together (Results and Discussion).

In this Unit, we shall be concentrating mainly on the Results section, although some
mention will be made to interpretation, a topic to be followed up in more detail in Unit VII.

In short, a typical Results section may follow this pattern

1. Results (data) (where are the results located in the text)

2. Most important findings
3. Comments

One interesting difference may lie between those texts which describe results and those
which try to re-describe them. In other words, to what extent will a researcher let the results
“speak for themselves” and to what extent will he or she intervene and do any or all of the
following:
1. generalise from the results
2. explain possible reasons for them
3. compare the results to those in other studies.

Activity 6.1
Read the texts which follow and identify those sentences in which the researchers are:

1. presenting results
2. commenting on the results in any of the ways mentioned above.

1.
RESULTS
Field Study
Respiration
(1)Microbial respiration, as measured by CO2 production, was increased by the addition of
ethanol with the labelled 15NO3 solution in the three depths above 120 cm when compared
with the labelled 15NO3 solution minus the ethanol (Fig.1). (2)At 120 cm, there was no
significant difference in CO2 concentration between the plus and minus ethanol treatments
(P  0.05). (3)Soil respiration increased with time and reached a maximum after 72 h.
(4)However, a decrease in respiration was observed with increasing depth in the soil
profile. (5)Production of CO2 from the soil surface was only slightly above background
levels with no difference between the ethanol and control treatments (data not shown).
(Water Quality: Potential for Bioremediation of High Nitrate Irrigation Water Via
Dentrification, KL Weier et al., Journal of Environmental Quality, 23:105-110 (1994).)

32
[Note: the Results section continues with paragraphs devoted to Labelled Gases and nitrous
Oxide Emissions before going on to the results of the Laboratory Study

2. (1)The results are shown in Figures 1-9. (2)All the methods track the modulated swing
very well, and only minor deviations have been observed in all scenarios with harmonic
levels up to 25 percent. (3)More pronounced was the noise sensitivity, especially when the
modified zero crossing method (Figures 1-4) was used. (4)This method is very good for
frequency measurements of stationary waveforms, but cannot cope with large amounts of
noise in the signal. (5)Both DFT with least squares fitting and demodulation techniques
successfully deal with large amounts of noise and harmonics, and could be used for
dynamic frequency tracking under very difficult conditions. (6)As expected, longer data
sets used for fitting in the DFT method produce better immunity to noise and harmonics.
(7)The best performance was obtained with 1 cycle DFT window for phasor calculation and
8 sample long window for quadratic polynomial fitting (Figures 5-8). (8)It should be noted
that ½ cycle phase shift observed on all estimates based on DFT is due to one cycle
measurement window - the estimates correspond to the middle of the measurement
window, but they can be calculated only when the whole window is available.
(9)Some improvement is possible by using advanced extrapolation techniques, which is the
subject of ongoing research. (10)The demonstration technique (Figure 9) does not require
long measurement windows and reaches excellent performance with a careful design of
filters. (11)The accuracy of the estimates is practically independent of the number of
harmonics and their magnitudes.
(Frequency Tracking in Power Networks in the Presence of Harmonics, Miroslav M.
Begovic et al., IEEE transactions on Power Delivery, Vol. 8, No. 2, April, 1993)
3. (1)After 4-5 days the biofilm consisted of cell clusters of 150-200 of 150-200 m
thickness separated by interstitial voids, the void fraction (of the substratum area) was
approximately 50%. (2)The substratum in between the cell clusters was covered with a 10-
20 m thick base film.
(3)Oxygen concentration profilers were measured at various bulk liquid flow rates.
(4)When the flow velocity was low (0.0078 m/s) the oxygen profiles above the void and
cell cluster areas were the same (Fig. 1A and B). (5)The strongest gradients occurred
between the depths of 120 and 230 m. (6)The vertical position of the mass transfer
boundary layer (determined graphically, see Fig. 1A and B) was 240 m from the
substratum. (7)The thickness of the mass transfer boundary layer (h) was approximately
220 m in the voids and 70 m above the cell clusters. (8)As the bulk velocity was
increased the h decreased above voids and cell clusters. (Fig. 2) (9)However, the h above
the voids decreased more rapidly than the h above the cell clusters and they were both
approximately 50 m at a bulk velocity of 0.04 m/s. (10)Increasing the bulk velocity
further caused the h above voids and cell clusters to become asympototic to the x-axis as h
approached zero.
(Liquid Flow and Mass Transport in Heterogeneous Biofilms, Dirk de Beer et al., Water
Research, Vol.30, No.11, 1996)

4.

33
 III. SIMULATION RESULTS

(1)A 3 hp, 220 V, 4 pole induction motor with the controller has been simulated. (2)The
parameters of the motor are given in Table 1. (3)A switching time of 50 secs is used in the
simulation. (4)A constant DC voltage of 225 volts is used as the input to the inverter.
(5)The command torque and command flux used are 11.9 N-m (rated torque) and 0.234 Wb
respectively. (6)Figs. 8 (a) and (b) show the torque and the stator flux responses of the
system during startup for the conventional DSC and the fuzzy controller. (7)The response
of the fuzzy controller is faster than the conventional DSC. (8)In the fuzzy controller the
initial stator flux error is very large. (9)Thus the controller chooses the states giving a
higher increase in the flux. (10)Once the flux becomes small, the controller chooses the
states giving a faster increase in torque. (11)Fig. 9 shows the response of the system for a
step change in torque from 11.9 N-m to 5 N-m, keeping the flux command constant.
(12)The response of the fuzzy controller is faster than the conventional controller.
(taken from Fuzzy Controller for Inverter Fed Induction Machines, Sayeed A. Mir et al.,
IEEE Transactions on Industry Applications, Vol. 30, No.1, January/February, 1994)

5.
SIMULATION RESULTS
(1)We performed simulations on small sets of cities taken from Tank and Hopfield (1985)
and from Durbin and Willshaw (1987).
(2)The algorithm described above is the result of several variations on a basic idea derived
from Kohonen’s self organization principles (1984). (3)The one we propose gives
satisfactory solutions in all cases. (4)Only the parameter  characterizing gain decrease has
to be adjusted, and the results obtained are not particularly sensitive to it. (5)Unlike
simulated annealing with temperature decrease, there is no need to use very low values of
, as no better solution would be obtained.
(6)Figure 2 shows two representative histograms of tour lengths obtained with different
initial permutations of the 30 cities used by Tank and Hopfield (see Figure 1). (7)It
demonstrates that a low value of the parameter  gives a better average, but a high value
has the advantage that the optimum is sometimes reached; it also gives a good solution in
several tries in less time than one try with a low value of . (8)In either case, a good
average solution (less than 3% greater than optimum) may be obtained in 2 seconds on
classical hardware.
(9)We compared our self organization method with the elastic net method presented by
Durbin and Willshaw. (10)The results are displayed in Table 1 for their five sets of 50 cities
(Figure 3) and show similar characteristics to those above. (11)On the average, our
approach is equivalent. (12)However, the possibility of starting with a random order of the
cities gives us some chance of getting better results. (13)Two particular cases involving
known optima with larger numbers of cities have helped us in our preliminary evaluation.
(14)They are given in Figure 4.
(15)A solution found by our algorithm for a set of 1000 cities is presented in Figure 5.

Activity 6.2
Read over the 5 texts again and underline all the main finite verbs. What do you notice
about their use? What tenses are used? What other verbal elements are present?

34
Activity 6.3
The pattern of Results-Major findings-Comments is not necessarily linear. Check over the
texts again and decide if the author/s highlight the major findings and perhaps comment on
them after each result or all together at the end. Also, are what we could call the “stronger”
results mentioned first and the “weaker” ones last?

Activity 6.4
Particularly in studies which present some kind of quantitative results, it is common to find
the author/s comparing an initial state with the final one, one set of results with another, one
study with another and so on. Linguistically, an author may choose any of the following
devices to achieve this.

(a) using a noun (increase, decrease, rise, fall)

(b) using a verb (increased, decreased, rose, fell)
(c) using a comparative or superlative adjective (higher, lower, greater, smaller,
the greatest, the smallest).

How many examples of these devices can you find in the 5 texts?

Activity 6.5:
Take your text, analyse the results section in terms of its content (results-findings-
comments) and form (verb tenses-comparative/superlative forms).

UNIT VII: DISCUSSION

This unit continues naturally on from Unit VI in the sense that it includes some elements of
the Results and Discussion sections found in research articles, together with additional
comments as to the Conclusions section which invariable follows it.
As we commented in the Unit on Introductions, such sections tend to “move” from the
general to the particular. The opposite is generally true of results and discussion sections.

Some of the following moves may be present:

1. Background information: (Original Hypothesis) A recapitulation of main points,

highlighting theoretical information or reminding the reader of technical information.

2. Statement of Results: (Findings) An obligatory move, which may be cyclical (repeated

at intervals throughout the section). Probably, the strongest claims will be mentioned first.

3. (Un)expected outcome: Writer’s comments as to whether the results were expected or

not. Perhaps more common in social and “human” sciences.

35
4. Reference to previous research: A common move. Two main sub-types: reference for
purposes of comparison with present research and references for purposes of providing
support for present research.

5. Explanation: Especially used if the results are “surprising”.

6. Exemplification: To support the explanation.

7. Deduction and Hypothesis: Used to make a claim about the generalizability of some or
all of the reported results.

8. Recommendation: The writer advocates the need for further research or makes
suggestions about possible lines of future investigation.

However, it must be remembered that journals respect a certain format which will
determine how a contributor structures the article in these sections. Some may be divided
into clearly defined parts such as Results - Discussion - Conclusions, other typically
combine the first two and have a separate Conclusions sections, which sometimes only
summarises in the simplest way possible the major findings and their possible
applications/implications etc.

Activity 7.1
Read the following extracts and work out what moves are included in the Discussion and/or
Conclusions sections.

1.
Title: A methodology for solving multi-objective simulation-optimization problems (Radi
Teleb & F.Azadivar, European Journal of Operational Research 72 (1994).
Abstract: For many practical and industrial optimization problems where some or all of the
system components are stochastic, the objective functions cannot be represented
analytically. Due to the difficulties involved in the analytical expressions, simulation may
be the most effective means of studying these complex systems. Furthermore, many of the
problems are characterized by the presence of multiple and conflicting objectives. The goal
of this paper is to introduce a new methodology through an interactive algorithm for
solving this multi-objective simulation optimization problem.

Outline of the article:

Introduction
Mathematical formulation of the problem
Review of related work
The methodology: Concepts and definitions
Algorithm
Examples of applications

36
Conclusions
(1)This work presents a new approach for multicriteria optimization of systems that are
evaluated through computer simulation. (2)The methodology suggests a new way of
defining a compromise solution in terms of the maximum likelihood that a solution will
contain the ideal solution consisting of optimum values for all objectives.
(3)This methodology can be applied to problems where the objective functions and the
stochastic constraints are distributed reasonably close to a normal distribution. (4)Although
average responses obtained from simulations often satisfy normality conditions, the user
needs to be aware of this assumption.
(5)It is not claimed that the complex search used here is suitable for all problems, at all
times. (6)Neither is it implied that the optimality definition used here satisfies the utility
functions of all decision makers. (7)The method is presented as one possible way of
multiple criteria optimization of complex stochastic systems.
(8)An automated algorithm is developed that can be employed when linked with a
computer simulation program. (9)This algorithm allows an interactive approach where the
user interferes with the findings of the algorithm. (10)Although potentially the algorithm
can be applied to any system with stochastic multicriteria responses, it is most suitable for
systems modelled by computer simulation. (11)This is true because of the ease of running
simulation experiments instantaneously and obtaining enough samples of the responses on
line to be able to perform the required statistical analysis.
(12)The paper also demonstrated the application of the algorithm to two problems.

2.
Title: Tower Surge Impedance (M. A. Sargent & M. Darveniza, IEEE Transactions on
Power Apparatus and Systems, May 1969)
Abstract
Introduction
This paper describes theoretical and experimental investigations of the surge response of a
transmission line tower, undertaken to clarify its representation in the calculation of the
lightning performance of a transmission line. These investigations comprised geometric
model studies and theoretical analyses of the response of simple tower forms to surge
currents of various waveshapes using electromagnetic field theory.
Geometric Model Studies
Theoretical Analyses
Discussion
(1) The use of a geometric model to determine tower surge impedance is considered to be a
theoretically rigorous and convenient technique of measuring this parameter. (2)Results of
measurements on these models, for a range of towers, indicate that the tower surge
impedance may be conveniently and accurately represented as a constant impedance in
traveling wave analysis. (3)This is demonstrated by the good correlation achieved for all
towers and waveshapes considered between measured model potentials and those
calculated assuming constant tower surge impedance, when both are examined as a
function of time. (4)The magnitudes of surge impedance determined on the model are in
good agreement with values calculated using Jordan’s formula, but differ greatly from those
derived by the Wagner and Hileman formula (2).

37
(5)Analyses of the surge response of simple tower forms have been performed using
electromagnetic field theory as the fundamental approach. (6)The conical representation
proposed for steel lattice towers is shown to have a surge impedance which is time
invariant, and hence independent of current waveshape. (7)Further, the magnitude of the
surge impedances determined in the analysis for the conical tower (130-150 ohms) is in
good agreement with the measured values (120-150 ohms) obtained using the geometric
models. (8)This close correlation between measured and calculated tower surge impedance,
both in magnitude and time invariance, supports both the representation of a tower by a
constant impedance and the use of the conical representation of transmission line towers,
instead of the usual cylindrical approximation.
.......................

3.
Title: Self-Organizing Feature Maps and the Travelling Salesman Problem (B. Angéniol et
al., Neural Networks, Vol.1, 1988)

Abstract: Based on Kohenen’s work on self-organizing feature maps, we derive an

algorithm for solving the classical Travelling Salesman Problem. Given a set of cities
defined by their positions in the plane, an evolving population of cells, featuring
duplication and selection, iteratively organizes towards a quasi-optimal solution.
Simulations on sequential machines have been carried out for problems with a thousand
cities. The algorithm naturally lens itself to implementation on a network of interconnected,
analog processing elements.
Introduction
Previous work on TSP
Algorithm Description
Simulation Results
(1)We performed simulations on small sets of cities taken from Tank and Hopfield (1985)
and from Durbin and Willshaw (1987).
(2)The algorithm described above is the result of several variations on a basic idea derived
from Kohonen’s self organization principles (1984). (3)The one we propose gives
satisfactory solutions in all cases. (4)Only the parameter  characterizing gain decrease has
to be adjusted, and the results obtained are not particularly sensitive to it. (5)Unlike
simulated annealing with temperature decrease, there is no need to use very low values of
, as no better solution would be obtained.
(6)Figure 2 shows two representative histograms of tour lengths obtained with different
initial permutations of the 30 cities used by Tank and Hopfield (see Figure 1). (7)It
demonstrates that a low value of the parameter  gives a better average, but a high value
has the advantage that the optimum is sometimes reached; it also gives a good solution in
several tries in less time than one try with a low value of . (8)In either case, a good
average solution (less than 3% greater than optimum) may be obtained in 2 seconds on
classical hardware.
(9)We compared our self organization method with the elastic net method presented by
Durbin and Willshaw. (10)The results are displayed in Table 1 for their five sets of 50 cities
(Figure 3) and show similar characteristics to those above. (11)On the average, our
approach is equivalent. (12)However, the possibility of starting with a random order of the

38
cities gives us some chance of getting better results. (13)Two particular cases involving
known optima with larger numbers of cities have helped us in our preliminary evaluation.
(14)They are given in Figure 4.
(15)A solution found by our algorithm for a set of 1000 cities is presented in Figure 5.
Conclusion
(16)Our approach to optimization problems through self-organizing feature-maps appears
very promising due to the following features. (17)First, the total number of nodes and
connections in a connectionist parallel implementation will be proportional only to the
number of cities in the problem, thus scaling very well with the problem size: hardware
based on analog technology, in a way similar to previous proposals (Hopfield, 1985), could
efficiently handle a large number of cities or variables. (18)Second, only one parameter,
which directly controls the total number of iterations, has to be tuned. (19)Finally, typical
values of this parameter ensure a good, near-optimum solution in a reasonable time for the
simulations we made. (20)Theoretical work is now necessary to support these experimental
results, to provide a foundation for further developments. (21)We plan to generalize his
method to other optimization problems by varying the organizing topology.

4.
Title: Fuzzy Controller for Inverter Fed Induction Machines (S. A. Mir et al., IEEE
Transactions on Industry Applications, January-February, 1994)

Abstract
Introduction: This paper presents a fuzzy controller for stator direct self control of an
induction machine. Fig. 4 shows a block diagram of proposed fuzzy controller which
modifies the direct self controller by incorporating fuzzy logic into it. Also, a fuzzy
resistance estimator is discussed to estimate the change in the stator resistance. The
change in the steady state value of stator current for a constant torque and flux command is
used to change the value of stator resistance used by the controller to match the machine
resistance.
The Fuzzy Controller
Simulation Results
Correction of Error Caused by Change in Stator Resistance

Conclusion
(1)A fuzzy logic controller for a direct self control of an induction machine has been
presented in this paper. (2)A response faster than the conventional DSC during startup and
during a step change in torque is achieved. (3)The performance has been tested by
simulations. (4)Also, a fuzzy resistance estimator has been proposed to correct the stator
resistance used by the controller to match the actual stator resistance of the machine. (5)The
stator resistance used by the controller follows the actual stator resistance of the machine.
(6)The algorithms are constructed upon experience with the motor drive system. (7)The
performance of fuzzy resistance estimator has been tested by simulation.

Activity 7.2: Library work

In Unit VI, you were asked to study the results sections of your text from the point of view
of their content and form. Look over them again and relate the structure and content of the

39
Results and Discussion and/or Conclusions sections with the original hypothesis or purpose
of the work.

40