On Subject and Theme A Discourse Functional Perspe... - (9. Themes, Methods of Development, and Texts)

9
Themes, Methods of Development, and Texts*

Peter H. Fries
Central Michigan University
and Hangzhou University
1. Introduction
This paper explores the notion of Theme as it is conceptualized in the
systemic functional model. The notion of Theme was clearly articulated by
Mathesius as early as 1939 and has been developed by members of the
Prague school since then. In the 1960s, M.A.K. Halliday, inuenced by
work within the Prague school, integrated a similar notion (which he called
Theme) into the systemic functional model. Fries (1981) discussed the
Copyright 1995. John Benjamins Publishing Company. All rights reserved.
similarities and dierences between the Prague school notion of Theme and
the concept of Theme which was used in systemic functional theory. Much
has been written since the sixties either supporting or disputing the various
claims associated with Theme in systemic functional grammar. A large body
of work uses Theme as part of the descriptive apparatus (Martin 1992, Ragan
1987, Rothery 1990, Plum 1988, Vande Kopple 1991 etc.). At the same
time, a number of authors are quite critical of Theme. (See Brown and Yule
1981; Chafe 1976; Gundel 1977; Hudson 1986; Huddleston 1988, 1991,
1992 for examples.) Clearly, the concept of Theme stands in need of clari-
cation and of further exploration. I intend to explore Theme in this paper by
examining four hypotheses which concern the relation between Themes and
texts1.
* This paper was written while on sabbatical leave from Central Michigan University. I would
like to thank Ruqaiya Hasan for detailed comments on an earlier version of this paper.
1. One issue concerning Theme is the relation between Theme and Subject. Such a discussion
is outside the scope of this paper.
On Subject and Theme : A discourse functional perspective, edited by Ruqaiya Hasan, and Peter H.
Fries, John Benjamins Publishing Company, 1995. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/coventry/detail.action?docID=
Created from coventry on 2017-12-18 11:02:16.
318 fries
2. Review of previous work

Systemic grammarians generally base their descriptions of Theme on the
classic descriptions provided by Halliday (e.g. 1967, 1970 and 1985). In
these works Halliday described the Theme of the clause in the following
terms:
The theme is what is being talked about, the point of departure for the clause as
a message (1967:212)
The theme is the element which serves as the point of departure of the message;
it is that with which the clause is concerned. (1985:38)
The English clause consists of a theme and a rheme [the theme] is as it were
the peg on which the message is hung, The theme of the clause is the element
which, in English, is put in rst position; [1970:161]
These passages dene Theme as the point of departure of the message, the
peg on which the message is hung or what the message is about, and they
say that Theme can be recognized in English by the fact that it occurs rst.
Elsewhere, Halliday has said that the element of structure Theme may exist
at several ranks, including group rank and the clause complex, in addition
to the clause rank. A number of works (e.g. Bcklund 1990, Fries 1981, in
press a) have examined Themes in t-units2. Hallidays wording what the
clause is about has often been assumed to equate Theme in systemic func-
tional grammar with what others called topic, and a number of articles have
devoted considerable space to demonstrating that what occurs rst in clauses
is often not the topic. (See Gundel 1977; Downing 1991; and Huddleston
1988 for examples of this interpretation.) I also showed (Fries 1981, in press
b) that Theme is not topic (or given or even necessarily nominal). Rather,
Theme functions as an orienter to the message. It orients the listener/reader
to the message that is about to be perceived and provides a framework for
the interpretation of that message3.
In his early presentations, Halliday generally discussed isolated sen-
2. A T-unit is a clause complex which contains one main independent clause together with
all the hypotactic clauses which are dependent on it. For an early description of T-units, see
Hunt (1965).
3. See Bcklund (1989) and Winter (1977:475) for similar wordings.

themes, development and texts 319
tences as examples. I found that his descriptions were persuasive, but felt
that his case could be made stronger by considering the thematic contribution
to texts considered as wholes. If Theme is a meaningful element on the level
of clause or clause complex, then we should nd that the kinds of meanings
that are made thematic would vary depending on the purposes of the writers.
Further, we should be able to manipulate readers and listeners reactions to
texts by changing the content of the Themes of those texts in much the same
way that we can manipulate reader and listener reactions to texts by changing
the words of the text (e.g. by changing and to but in certain places). With
this end in view, I investigated a number of short (one and two paragraph)
texts to see whether thematic content within these texts related to the pur-
poses of the authors and the perceptions of readers. The results of that study
were presented in Fries (1981).
In that paper, I hypothesized that:
(i) dierent patterns of Thematic progression correlate with dierent
genres, i.e. patterns of thematic progression do not occur randomly but
are sensitive to genre; and
(ii) the experiential content of Themes correlates with what is perceived to
be the method of development of a text or text segment 4.

Since Fries (1981), two additional hypotheses have been articulated:
(iii) the experiential content of Themes correlates with dierent genres, and
(iv) the experiential content of the Themes of a text correlates with dierent
generic elements of structure within a text.
Clearly all four hypotheses mentioned here concern the relation between
some aspect of the information which is used as Theme in a text and some
aspect of the text as a whole. However these hypotheses relate thematic
content to slightly dierent aspects of texts. The work published on Theme
since my initial paper has generally supported the four hypotheses listed
above. However this work has also shown that the picture is complicated.
Things are not as simple as the above wording of the hypotheses would lead
one to believe.
4. A closely related hypothesis presented in Fries (1981) is indirectly relevant to this paper:
information which is perceived to constitute the point of a text or text segment is regularly
found within the Rhemes. In other words, information which is perceived to present the point
of a text segment would not be found in the Themes.
320 fries
2.1 Hypothesis 1
The rst hypothesis concerns the relation of dierent thematic progressions
to dierent genres. The notion of thematic progression, derived from the
work of Dane (1974) and others, concerns the ways that texts develop the
ideas they present. More specically thematic progression concerns where
Themes come from how they relate to other Themes and Rhemes of the
text. Patterns of thematic progression are formed by a systematic relation
between the Theme-Rheme selections and experiential selections in a text.
Thematic progression may be investigated by exploring the cohesive ties
which occur within the various Themes within a text and the locations of the
items the cohesive ties presume. Thus, given a Theme, what are the cohesive
items within that Theme and where are those cohesive items resolved? Dane
described a number of typical Thematic progressions, including the three
diagrammed in Figure 1.
The rst type of Thematic progression in Figure 1 could be called
linear thematic progression. In linear thematic progression, the content of
the Theme of a second sentence (Theme 2) derives from the content of the
previous Rheme (Rheme 1), the content of Theme 3 derives from Rheme 2
etc. The second type of thematic progression could be called Theme itera-
tion. In Theme iteration, the same Theme enters into relation with a number
of dierent Rhemes. The result of this type of Thematic progression is that
the Themes in the text constitute a chain of (typically) co-referential items
which extends through a sequence of sentences or clauses. The third type of
thematic progression might be called a progression with derived Themes. In
this case, the passage as a whole concerns a single general notion, and the
Themes of the various constituent clauses all derive from that general notion,
but are not identical to one another.5 Thus, an obituary might use Themes
which refer to the person who died, services, the funeral, burial, etc. Clearly
these items are not coreferential, but they can all be seen to relate to the
situation as a whole.
Subsequent work on Thematic progression has only weakly supported
hypothesis 1. For example, Gill Francis (1989, 1990) examined articles
5. Cloran (this volume) points out that one must also consider patterns of progression in the
Rheme, adding at least Rheme Rheme progressions. See also her characterisation of the
thematic progression patterns.
79m x 122m
Figure 1: Three patterns of Thematic progression

according to Dane
which were found in three contexts in two newspapers. She examined the
clause Themes of front-page news reports, editorial comments, and letters
to the editor which expressed some sort of complaint. She took the dierent
placement in the newspapers to indicate that the articles belonged to dierent
genres, and assigned news reports to the news report genre, editorials to
analytical exposition (persuading that), and letters of complaint to the genre
of hortatory exposition (persuading to). Her results supported hypothesis 1
322 fries
in that she found that the Thematic progressions used in the news reports
diered from the Thematic progressions used in the other two genres. In
particular, news reports contained many sequences which were reminiscent
of pattern two in Figure 1 (Theme iteration), while the editorials and letters
to the editor which she examined contained fewer such instances. As she
said:
It is clear that in News, far more Themes are involved in chains, the chains are
longer, and the average number of tokens per chain is higher than in the expository
genres. (Francis 1988: 212)
Further, she found that the chains in which the Themes in news reports
participated diered from the chains she found in the expository genre. The
Themes within the News genre belonged largely to identity chains, while the
Themes of the expository genre made far greater use of similarity chains6.
An identity chain consists of a series of language elements which refer to the
same thing (or event). A similarity chain is a series of language elements
whose referents are non-identical, but belong to the same or related classes.
(See Hasan, 1984: 205-206 and 1989: 84-85, for more extended descrip-
tions.) However, Francis also found that news reports contained many
exceptions to this second pattern of thematic progression. Further, she
reports that the picture was even less clear when she looked at the texts from
the two expository genres. That is, the texts from the expository genres did
not seem to exhibit a clear thematic progression, and certainly were not
distinguishable from one another. In Fries (1981) I had predicted that
argumentative expository genre should use Danes rst pattern of thematic
development (linear thematic progression). Francis found some examples
of this pattern of development in her expository texts, but these texts also
contained many unexplainable exceptions to this pattern. The exceptions
were so varied that she was unable to discover any general pattern of the-
matic progression in her expository texts. She described the situation in the
following terms:
There is an extraordinary diversity here and it is dicult to see a Theme-Rheme
pattern emerging, even at paragraph level. (Francis 1989: 215)
6. Of course identity chains and similarity chains may be found within any part of a clause
or sentence (Theme or Rheme). For the purposes here, Francis explored only those chains
which contained at least one member which occurred within the Theme of a clause.
Bcklunds (1990) discussion of eight telephone conversations from the

London-Lund Corpus makes a similar point. While Bcklund nds 61
instances of Theme iteration (Danes second pattern) most of these in-
stances are found in sequences of two or three identical Themes. She nds
40 instances of Danes rst pattern, but all of these instances consist of
minimal chains of two links. In other words, both Francis and Bcklund nd
at least some genre which seem not to have a clear pattern of thematic
progression. Both Francis and Bcklund seem to be looking for uninterrupted
thematic progressions and make the point (in dierent ways) that one needs
to explore chains which are interrupted by other material. However, it seems
unreasonable to expect that every Theme of a text, particularly a long text,
should t into a single pattern of thematic progression. Most texts achieve
several purposes, and one should expect the grammatical structures used in
these texts (including choice of Themes) to change as one or another purpose
is addressed. Francis suggests that even chains with relatively few tokens
may range over large portions of text in news reports, forming one device
which news writers use to maintain coherence over long stretches of text.
Hypothesis 4 suggests that the intermittent recurrence of specic information
as Theme (e.g. as described in Francis) might signal some aspect of text

structure. (This topic will be discussed again later.)
It is clear that in its present form, hypothesis 1 does not account for all
the data. It is possible that it may be improved by integrating a view of text
structure (or invoking the notion of information types, following Longacre
1981, 1989, 1990).
2.2 Hypothesis 2
The second hypothesis concerns the relation between the experiential content
of the Themes of a text and readers and listeners interpretations of that text.
Two texts may say roughly the same thing but develop their ideas in dier-
ent ways. The way in which a text develops its ideas can be called the
method of development of the text. The method of development of a text
aects the reactions of its listeners and readers. There is no requirement that
a good text develop its ideas in a single self-consistent manner. However,
some texts do so. The second hypothesis claims that if a text is perceived to
have a single simple method of development, then the Themes of the clauses
and clause complexes of the text will be seen to express meanings which
324 fries
relate to that method of development. Most typically, texts which are per-
ceived to develop their ideas in a single simple way (i.e. which are perceived
to use a single simple method of development) base that development on just
a few experiential meanings being presented in the text. Hypothesis 2
predicts that the texts which take that approach will use the experiential
content of the Themes to construct that development. A corollary of this
hypothesis is that if we change the experiential content of the Themes of a
text, we will change readers perceptions of how the ideas in that text are
developed. Thus, this aspect of hypothesis 2 should be testable using stan-
dard psychological testing techniques measuring readers interpretations of
texts.However, no one has yet carried out such an experiment systematically.
I should point out that the notion of method of development is not a
structural idea but a semantic one. Dierent texts may be perceived to
express single methods of development to varying degrees. Or, to put it
another way, some texts develop their ideas in simple ways, while others
develop their ideas in quite complicated ways. Hypothesis 2 predicts that the
perceived simplicity or complexity of the development of the ideas in a text
will correlate with the degree to which the experiential content of the
Themes of the text may be seen to be derived from a limited set of semantic
elds. Some studies, such as Thompson (1985) and Francis (1989, 1990),
do not nd texts with long series of Themes which derive from a single
semantic eld. This nding in itself does not contradict the hypothesis, since
it focuses only on the formal characteristic of the texts. Such a nding would
only contradict this hypothesis if the texts were also perceived (by listeners
or readers) to have a single simple method of development. While Thompson
and Francis do not discuss the semantic impression conveyed by the texts
being analyzed, the examples which they provide show that not only is it the
case that the experiential content of the Themes of the component clauses
and clause complexes does not derive from a single semantic eld, but also
these texts develop their ideas in relatively complex ways. Hence these
studies may in fact be taken to partially support hypothesis 2.
Further partial support for hypothesis 2 is supplied by the many studies
of individual texts or small groups of texts in which the experiential content
of the Themes of the clauses of the text is examined and tied to the interpre-
tation of the passage. These studies (For example Benson, Greaves, and
Stillar 1992, Halliday 1993, Martin 1986, 1989, and 1992, and Ragan 1987),
report a correlation between the experiential content of the Themes of the
clauses and the impression produced by the texts examined. Ragan, for
example, studied directions written by native and nonnative speakers, and
found that sets of directions contained three major elements of structure: one,
which was obligatory, described the procedure to be done, a second (op-
tional) element of structure oriented the reader to the task, while the third
(again optional) element of structure recapitulated and concluded what was
said earlier. The descriptions of the procedure to be done always used the
processes to be done (the actions) as method of development. This correlated
with the fact that the Themes of the descriptions of the procedures regularly
contained simple Themes which referred to processes. Thus, for example,
the procedure portion of the text produced by student mp1av-3 contained
14 non-embedded clauses. The Topical Themes of 11 of these clauses
referred to processes.
Similarly, Benson et al compared two critical reviews of a lm which
conveyed radically dierent impressions. They found that, among other
things, the methods of development of the two texts diered. One review
seemed to develop its point very simply, focusing on the lm as a product,
and that simple development correlated with the fact that six of eight topical
Themes refer to the lm itself, while the remaining two Themes refer to
characters portrayed in the lm. The second review took a more complex
approach to the task and presented the lm as a process in which many
people participated. This complexity in impression correlated with more
complex semantic patterning in the experiential content of the Themes of the
clauses of the text. Only one of 18 clause Themes referred to the lm as a
whole, and even this Theme also included the directors name. Other clause
Themes referred to the script and script writer, characters other than romantic
leads, actors, and costumes. Celebrities of the period represented in the lm
are separated from their portrayal in the lm in several Themes, etc. In other
words, Benson et als results follow the predictions of the second hypothesis.
2.3 Hypothesis 3
The third hypothesis predicts that the experiential content of the Themes in
a text is sensitive to dierent genres. At one level, this may appear quite self-
evident. For example, we expect tour guides to contain many references to
spatial location. Therefore, even if these references to spatial location were
to occur randomly in the clauses and sentences of a tour guide, we would
326 fries
still expect a relatively larger number of references to spatial location in the

clause and sentence Themes of such texts. However, this hypothesis predicts
that meanings are not randomly distributed in dierent positions in the
sentences and clauses of a text. A basic purpose of tour guide texts is to
present and explain the various sights to the reader. References to spatial
location are typically present to orient the reader to the location of these
sights. Therefore, because of the purposes of tour guide texts, references to
spatial location will play an orienting role. This orienting role will lead
authors of tour guide texts to place a greater proportion of their references
to spatial location within the Themes of the text. On the other hand, in
narratives, references to spatial location in clauses which refer to event line
processes7 typically play a much more focal role. Due to this dierent textual
role, references to spatial location in event-line processes will be less likely
to occur thematically. Thus, this hypothesis cannot be explored with simple
raw counts of how many Themes refer to spatial location. Rather, when
comparing texts we must examine all references to a type of meaning (such
as spatial location) in the texts and then compare the proportions of thematic
versus non-thematic occurrences.
The third hypothesis has been explored in greater detail than the second.
Indeed, several papers over the last few years (e.g. Berry 1987, Bcklund,
1990, Francis 1989 and 1990, and Wang Ling 1991 and 1992) have focused
on exactly this issue. It is of interest to nd that the results have varied
depending on which genres were explored. Franciss (1989 and 1990)
comparison of Themes in news reports, editorials and letters of complaint
found that the Themes used in the news reports diered from the Themes
used in the other two genres. She found that The typical and most predomi-
nant Themes in News are material and verbal, while far fewer relational
participants/processes are selected as point of departure for the message
(1990: 53). (I believe this result could be paraphrased as the Themes of the
clauses in news reports referred much more often to verbal and material
processes or to participants in verbal and material processes than they did
to relational process or participants in relational processes. More simply put,
News contained far more material or verbal processes than relational pro-
cesses, and as a result references in the clause Themes to portions of these
7. The term event-line processes refers to the description of events in a story line. (See
Longacre 1981, 1989, 1990.)
processes (the participants), or to the processes themselves were more

frequent than references to relational processes or to participants in relational
processes.) By contrast, Editorials and Letters both thematize fewer
material processes, very few verbal processes and a much higher percentage
of relational processes. (Francis 1990:53)
Francis also investigated the sorts of information which were thematized
in the three genres. She investigated the distributions of four basic semantic
categories of Theme:
1) people/groups/institutions i.e. all humans individually and collectively; 2)
concrete things such as aeroplane and dress; 3) abstractions, which includes both
nominalized processes like inquiry , proposal, refusal, etc., and the more domes-
ticated abstract nouns like job, percentage, and law; and 4) time expressions.
(Francis 1990: 55).
A fth category called other consisted of processes or non-lexical
Themes. The distributions of these information types is given in Table 1
(adapted from Francis Figure 3).
Table 1: Distribution of typical lexis of Theme in Francis data

Information type News Editorials Letters

1) people etc. 107(=53%) 32(=22%) 52(=28%)
2) concrete things 18(=9%) 13(=9%) 12(=7%)
3) abstractions 43(=21%) 61(=41%) 62(=34%)
4) time 10 7 10
5) other 23 30 47
Total Themes 201 143 183
1) + 2) 125(= 62%) 45(= 31%) 64(= 35%)
Clearly News contains far more people and thing Themes than the other
two genres. By the same token, the two expository genres use abstractions
as Theme much more often than News does.
Francis also investigated the distribution of thematic equatives (What
the Duke gave to my aunt was that teapot.), predicated Themes (It was John
who wanted us to go.) or Embedded (rankshifted) clausal Themes (That
Caesar was dead was obvious to all.) and modifying clausal Themes (Far
from your leader (June 5) containing misstatements of fact and misunder-
standings, as Mr Bailey alleged, it was, in my view, very accurate). (Themes
328 fries
in these examples are indicated in bold). The Themes of Editorials and Letters
contained more examples of these constructions than did the Themes of
News. Table 2 (her Figure 5) presents the results of her analysis of her data.)
Table 2: Distribution of clausal Themes in Francis data

News Editorials Letters
embedded clausal Themes 0 2 2
thematic equatives 1 3 1
predicated Themes 1 3 2
modifying clausal Themes 4 9 11
Total 6 17 16
Of course the structures just listed may be regarded as simply grammatical

constructions. However, these grammatical constructions have semantic
consequences. Embedded clauses are means by which processes may be
made into participants. As a result they provide an index of the range of
phenomena which may serve as Theme. Does this range include or not
include processes and facts, etc.? Thematic equatives and predicated Themes
involve markedness. Typically this markedness is a markedness of focus or
of Theme (or both). In either case, this sort of markedness is associated
with degrees of intervention of the writer. (Francis 1990: 60) As a result,
thematic equatives and predicated Themes provide an index of how often
writers use some sort of marked textual structure. Since news reporting is
presented as objective while letters of complaint and editorials involve overt
authorial intervention, we should expect grammatical structures which
overtly encode marked focus to occur more frequently in letters of complaint
and in editorials than in News8.
8. While Francis describes the relative frequencies of those embedded clauses, nominaliza-
tions, etc. which occur thematically in her data, she does not indicate what proportion of all
nominalizations etc. are found thematic in her texts. As a result we cannot determine if the
frequencies which she reports arise directly from the overall frequencies of these construc-
tions in the three genres or whether the constructions are being used in dierent ways in the
three text types. For example, it is possible that the dierences between News and the other
genres arises because of the dierences in overall frequencies of these constructions in the
three genres. News may simply use many fewer nominalizations, etc., and therefore fewer
of these constructions are found thematic within the clauses of News.
Finally, it is a feature of the News reports which Francis studied that the
clause Themes of news reports contained a high proportion of lexical items.
Francis refers to this index as an index of lexical density (see also Ure 1971
and Halliday 1989 Chapter 5) and nds that Themes in editorials have an
over all average lexical density of 2.0, Themes in News reports 1.7 and
Themes in letters of complaint 1.5. Much of the high lexical density in News
reports was achieved through the use of apposition within the various
Themes. Such use of apposition was not characteristic of either letters to the
editor or of editorials.
In summary, Francis results supported hypothesis 3 only partially. The
Themes used in news reports diered clearly in content from the Themes of
the two expository genres. However, when Francis compared the Themes
of the Letters with those of the Editorials, she found that the overall averages
for each of the two genre did not dier very much. Further, the variation
between individual texts within each of the two expository genres was so
great that the dierences in the averages which she found did not seem
signicant. This nding conicted with the results that Martin (1986, 1989)
reported in which he found very marked dierences between the experiential
content of the clause Themes of hortatory and analytical exposition. In

commenting on the dierences between her results and Martins, Francis
(1990: 54) focuses on the small size of the samples being examined, con-
cluding that The variability from one text to another suggests that the size
of the sample is a major factor in determining ones ndings. Clearly she
is correct in this. Studies based on three or four instances of a genre can be
taken as suggestive at best. I would like to add, however, that the criteria
which are used to gather and make an initial sort of the data are also critical.
It is dangerous to assume too close an association of physical position (e.g.
in a news paper) with genre characteristics. While, for example, it may be
reasonable to assume that editorials dier in regular ways from news reports,
is it reasonable to assume that all editorials (or news reports) belong to the
same genre and that they are therefore uniform? Francis results would seem
to argue a negative answer. (See Hasan in press for a discussion of the
constitutive role of language in social interactions.)
In contrast to Franciss study, a number of other researchers have found
the experiential content of the clause Themes to correlate fairly strongly with
genre type. In a study of the use of place names in texts from four genres,
viz. a guide book, a travel brochure, a transcript of a coee party, and a tran-
330 fries
script of a committee meeting, Berry (1987) provided evidence which

suggests that place names have dierent distributions in the four genres. She
examined the rst 25 place names which occurred in each text and found the
distributions presented in Tables 3 and 4 (= Berrys Tables 2 and 3).
Table 3: Occurrences of place names in Subject,

Object, Complement, Adverbial
Coee Committee Travel Guide
party meeting Brochure book
Subject 0 3 12 16
Object 3 5 1 4
Complement 1 2 1 0
Adverbial 21 15 11 5
Table 4: Occurrences of place-names in Theme

with comparative gures for Subject
Coee Committee Travel Guide
party meeting Brochure book
Theme 2 3 10 14
Subject 0 3 12 16
Themes which are 2 0 4 0
not Subjects
Subjects which 0 0 6 2
are not Themes
The guide book and travel brochure clearly use place names as Subject
and as Theme much more often than the other two texts. In examining Table
3 it is important to remember that all Berrys texts had 25 instances of place
names. Thus, in the coee party text, 23 of the 25 occurrences of place
names were neither Theme nor Subject. Similarly, 22 of the place names
used in the committee meeting functioned neither as Theme nor as Subject.
On the other hand, 10 of the 25 place names in the travel brochure and 14
of the 25 place names in the guide book functioned as Theme. While the
numbers in this study are small, they are supported by related projects my
students have done. For example, a class project by Nie (1991) studied the
choice of Themes in a portion of a guide book to Hangzhou and found that
six of 14 clause Themes referred to spatial location. Examples of such
Themes included At the top of Jade Emperor Hill, half way up, From this
point on the road, etc. Another group of six Themes contained names of
places or portions of the sites, e.g. the well of the site, Phoenix Hill, the three
sitting Buddhas, etc. Since each of these refers to portions of sites, we may
think of these six Themes as at least implying spatial orientation. Thus,
twelve of the fourteen Themes in this section of the guide book provided
either direct or implied spatial orientation.
In a study which is part of her Masters Thesis at Peking University,
Wang Ling (1991) examined ve-page sections of six plays by Sam Shep-
pard, and compared the language of the text which is to be spoken by the
actors (the line text) with the text which provides stage directions and
descriptions, etc. (the scene text). She found that the Themes in the scene
text diered from the Themes of the line text. Scene text used a far higher
percentage of simple Themes than did line text. (e.g. Buried Child showed
a ratio of 28/29 =96%, vs. 74/163 = 45%). In other words, scene text uses
the experiential metafunction as the source for thematic material practically
exclusively. Within the simple Themes, scene text used a higher percentage
of marked Themes than did Line text. (Again, Buried Child showed the
following ratios 6/28 = 21% vs. 4/74 = 5%). While Wang Ling did not
describe the semantic nature of the marked Themes in scene text, her exam-
ples include:
(1) In the dark, the light of the lamp and the tv slowly brighten in the black
space.
(2) Up right is an old, dark green sofa with the stung coming out in spots.
Clearly spatial location constitutes an important point of departure for

scene text. Further, the Themes of the scene text contained very little varia-
tion in construction or in meanings when compared to Themes used in the
line text. Wang Ling found very few instances of clauses used as Themes,
and no instances of predicated Themes in the scene text. She does not give
gures for the occurrence of these constructions in the line text, but does say
that the line text contained many examples of these constructions.
The last study I report on here examined eight sample conversations
from the London Lund Corpus of Spoken English. In this study, Bcklund
found that in fully 51% of her data (216 t-units) the ideational Theme
refers to one of the speakers (I, you) or to a group including one of the
332 fries
speakers or both (we, you)9. Further, all but 19 of the 419 t-units have the
Subject as ideational Theme. Of the 19 marked Themes, 15 are Adverbial
clauses, 3 are temporal now, and one is from then. None of the texts studied
in the works reported on here show such a heavy concentration of Themes
which refer to participants in the conversation. However, a comment in
Berry (1987: 84) suggests that speakers in the coee party text which she
analyzed used personal Themes more than absolutely necessary.
Table 5 summarizes several features of the works which have just been
described and one which will be discussed in Section 2.4. Since the studies
were performed by dierent workers who measured dierent features, the
numbers are not all comparable. For example, only one of the ve authors
(Xiao) enquired into the eect of the generic structure on the phenomena
examined. Text lengths were described using word count (Francis) clause
count (Wang and Xiao) and t-unit count (Bcklund and Nie). Luckily
Francis also mentioned the number of Themes she examined, and this
number provides at least a minimum count of the clauses in her data. Table
5 is divided into two parts, Table 5a and 5b. Table 5a focuses on general
descriptions of the texts and on the grammatical structures which function
as Theme. Table 5b provides information on the use of dierent types of

meanings in the Themes of the various texts. Yes in the column for line
text under Wang indicates that Wang reports the occurrence of embedded
clausal Themes and predicated Themes, but provides no counts of her data.
Since Wang provides no specic information on the use of dierent types
of meanings in the Themes of scene text and line text, no column is provided
for her work in Table 5b. Berrys work was suciently dierent from the
others as to make it impossible to place it in Table 5 at all. Dashes in the
columns for the texts studied by Francis (see bottom of Table 5b) indicate
that Francis did not separate out locational meanings as a separate category,
and so these meanings may be included in the gures for her Other cate-
gory.
9. This result is supported by other researchers as well. Givn suggests that humans tend to
make humans the topic of their conversations. Berry (1987: 84) reported that in the casual
conversation she analyzed speakers have actually avoided assigning place-names to
grammatical subject even where this would have been topically appropriate.
Copyright 1995. John Benjamins Publishing Company. Al

Table 5A: Summary of properties of Themes reported in Francis, Wang, Bcklund, Xiao, and Nie
Francis Wang (Buried Child) Bcklund Xiao Nie
Genre type: News Reports Editorials Letters to Eds Scene Text Line Text Conversation Recipes Fables Guide book
Generic Elements Intro Procedure
Length of Text 2500 wds 2500 wds 2500 wds
201 Themes 143 Themes 183 Themes 29 cls 163 cls 419 T-units 8 cls 38 cls 39 cls 14 T-units
Grammatical Structures in Themes:
Simple Themes 28 (96%) 74 (45%) 182 (43%) 8 (100%) 35 (92%) 32 (82%) 14 (100%)
Marked Themes 6 (21%) 4 (5%) 19 (5%) 0 6 (16%) 5 (13%) 8 (57%)

Embedded Clausal 0 2 2 0 yes 0 0 0 0
Themes
Thematic
On Subject and Theme1 : A discourse
3 functional
1 perspective,
0 edited by Ruqaiya Hasan, and 0Peter H. 0 0 0
EquatiFries,
ves John Benjamins Publishing Company, 1995. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/coventry/detail.action?docID=
Created from
Predicated coventry1 on 2017-12-18
Themes 3 11:02:16.
2 0 yes 0 0 0 0
334 fries
Table 5B: Experiential content of Themes reported in Francis, Bcklund, Xiao and Nie
Francis Bcklund Xiao Nie
Genre type: News reports Editorials Letters Conversation Recipes Fables Guide book
Generic Structure: Intro Procedure
Retrospective Text few freq freq 0 0 0 0
References
Speaker/Hearer 216 (51%) 0 0 0 0
Major Text 107 (53%) 32 (22%) 52 (28%) 6 (75%) 5 (13%) 26 (67%)

Participant
Concrete things 19 (9%) 13 (9%) 12 (7%) 3 (7.7%) 6 (43%)
Abstract concept 43 (21%) 61 (41%) 62 (34%) 5 (13%)
Process 27 (71%)
Time 10 7 10 1 (13%) 3 (7.7%) 1 (7%)
Location 6 (43%)
Created
Other from coventry on232017-12-18 30
11:02:16. 47 1 (13%) 6 (16%) 3 (7.7%) 1 (7%)
2.4 Hypothesis 4
The fourth hypothesis concerns the relation between the experiential content
of the Themes of a text and the dierent elements of structure within that
text. This hypothesis has two forms. The rst version of hypothesis 4
(henceforth hypothesis 4a) assumes that as a text moves from one element
of structure to the next, the purposes to be accomplished will change. As a
result, the meanings expressed in each element of structure of the text will
change and therefore the language used in each will change. The dierences
which are found in the Themes of the clauses of the dierent elements of
structure are simply part of the dierences in language which are generally
encountered. (This version of hypothesis 4 can be regarded as a more
sensitive version of hypothesis 3.) Plum (1988 vol. 1: 283) put it this way:
we would expect a predominant realisation of Theme in a way which allows
for the potential realisation of character, i.e. by Themes which are potentially
co-referential, to correlate not only with narrative-type genres generally but also
specically with the crisis stages in narrative. Similarly, we would expect other
stages, such as an end stage, to correlate with a choice of Theme which is not at
all concerned with the realisation of character but instead by a Theme which
foregrounds its metatextual function by making anaphoric reference to events
already related in the text, i.e. by either extended or text reference, since it is part
of the meaning of an end stage to wrap up the text.
Hypothesis 4a can in fact be made more specic. As we saw above, the

element of structure Theme has been described in terms such as point of
departure, peg on which the message is hung, framework for the inter-
pretation of the message and orienter. By contrast, Rheme has been
described as the unmarked location of New information10 information
which is newsworthy and to which the listener should pay attention. Since
the descriptions of the meanings of Theme and New dier in important
ways, we can predict that while both Themes and Rhemes will change as
a text progresses through its various elements of structure the two will
respond to dierent forces and will contain dierent sorts of information.
In a study of written English, Fries (in press a) coined the term n-Rheme
(for New/Rheme) for the last constituent of a clause the constituent which
10. See Halliday (1967: 201-202; 1985: 274), and Chafe (1984: 437).
336 fries
is the unmarked location of New information, and made the following

hypothesis:
As we examine the text, we should keep in mind that the n-Rheme is the news-
worthy part of the clause, that is, the part of the clause that the writer wants the
reader to remember. As a result we should expect the content of the n-Rheme to
correlate with the goals of the text as a whole, the goals of the text segment within
those larger goals, and the goals of the sentence and the clause as well. On the
other hand, the Theme is the orienter to the message conveyed by the clause. It
tells the reader how to understand the news conveyed by the clause. As a result,
we should expect the choice of Thematic content usually to reect local concerns.
For example, if we are examining a text which has a problem solution structure,
we should expect the meanings to change as the text moves from the description
of the problem to the description of the solution. Both the Thematic content and
the n-Rhematic content should change. However, the content of the n-Rhemes
should be more obviously connected with the goals of each text portion. For
example, in the section which describes the problem, the n-Rhemes should have
an obvious connection with what is wrong, while in the section which describes
the solution, the n-Rhemes should have an obvious connection with what was
done to solve the problem. The Themes of the problem section, on the other hand,
might well concern dierent aspects of the item which is causing the problem (say
an engine which is not functioning properly), while the Themes of the solution
section might concern notions such as the relative temporal order of the actions
taken in solving the problem11.
Fries found a relation between the purposes of texts and the information
found in the Themes and the n-Rhemes of written advertisements (1992,
1993) and of a fundraising letter (in press a). However Fries work does not
relate the experiential content of the Themes and the n-Rhemes to the
generic structures of the texts examined. While the specic hypotheses
relating the information in the Themes and the n-Rhemes to generic structure
suggested in Hypothesis 4a have not been seriously investigated, a number
of studies support the general trends predicted by hypothesis 4a. Ragan
11. The relationship which holds between the hypothesis described here and the discussion
of the functions of Theme Rheme and Rheme Rheme progressions found in Cloran (this
volume) needs to be explored. Clearly the claims are similar in that they focus on the
importance of considering the information placed in the Rheme. They dier in that one
focuses on the experiential content of the Rheme and the role of that content in the text, while
the other focuses on the ties between the various Rhemes of a text and their relation to the
structure of the text. As a result of this dierence, it remains to be seen whether they make
compatible predictions.
(1987) reports that the instruction genre which he studied contained three
elements of structure: an optional orientation element, an obligatory proce-
dure element, and an optional recapitulation element. A very high proportion
of the clause Themes of the procedure element referred to processes in his
data, while the clause Themes of the other two elements of structure rarely
referred to processes. For example in one of his texts, 12 of 15 clauses in the
procedure element of structure referred to processes. By contrast, 6 of 9
clause Themes of the recapitulation referred to the design as a whole, and
only two clause Themes (one of which is in a branched clause which has no
overt subject) refer to processes.
A class project by my student Qun Xiao nds similar trends. She
examined Thematic choices in recipes with thematic choices in fables. She
found that the recipes had four elements of structure, an introduction, a list
of ingredients (not analyzed in her project, since the lists consisted exclu-
sively of nominal groups), a description of the procedure, and a summary.
In the introductions, 6 of 8 Themes mentioned the dish as a whole. In the
descriptions of the procedure, 27 of 33 Themes referred to actions involved
in preparing the food. Another 5 Themes referred to instruments involved
in the preparation, and 1 Theme expressed a Circumstance of Manner for the

process. The ve Themes of the Summaries were all implicit and referred
to the dish as a whole. By comparison two thirds of the Themes which
occurred in the fables (26 of 39 = 67%) referred to major characters in the
fables. Five Themes referred to abstract concepts. All the references to
abstract concepts occurred in the expression of the moral, a portion which
clearly constituted a separate element of structure in the tales.
In a much larger study, Plum (1988) found a correlation between clause
Themes and generic structure. He divided the topical Themes in the narra-
tives he examined into (potentially) participant chaining Themes and non-
participant chaining Themes, and found that roughly 90% of the clause
Themes in elements of structure such as Incident and Crisis used participant
chaining Themes, while less than 60% of the clause Themes in elements of
structure such as Reorientation, Interpretation and Abstract used participant
chaining Themes12.
A second version of hypothesis 4 (henceforth hypothesis 4b) claims that
12. Plum (1988) also found that the textual themes of the clauses of the dierent elements
of structure of his texts diered as well.
338 fries
as one moves from one element of structure within a text to the next, that
move will be signaled in the Themes of the clauses and clause complexes13.
More specically, the Theme of the rst clause of the new element of
structure of the text will indicate that a new element of structure has begun.
This hypothesis is too strong as worded here, since it is quite easy to locate
specic examples where the move to a new element of structure of a text is
not signaled in the Theme of the rst clause or clause complex in any
obvious manner. However, if we consider this statement to describe a
statistical tendency, then it may be more acceptable. I do not know of much
empirical work which explores this hypothesis. Prideaux and Hogan (1993)
provide some suggestive information that initial placement of subordinate
clauses may have a discourse marking function. They gathered oral and
written narratives in controlled conditions, made a structural analysis of
these narratives (focusing on locating the boundaries of episodes in those
narratives). They then examined all instances of subordinate clauses in the
narratives, and compared the placement in episodes of sentences which
began with the subordinate clauses (sc + mc) with the placement of sen-
tences which began with the main clause (mc + sc). They found that 35 of
37 instances of the order mc + sc in the oral narratives occurred after the

episode had been initiated (i.e. not as the rst sentence of the episode). By
contrast 23 of 34 instances of the order sc + mc in the oral narratives
occurred as the rst sentence of a new episode. Further, the 11 instances of
the sc + mc ordering which did not occur initially within their episode (a)
referred back to the rst mention of the episode, (b) signaled the end of an
episode, or (c) signaled an especially important or salient event. (1993: 405)
They found similar trends in their written narratives. Clearly this evidence
suggests that something special happens at the boundaries of new discourse
elements. Of course, while this evidence is suggestive, it does not directly
address the issue of whether the experiential content of the clause complex
13. Another related hypothesis which needs to be explored is that the move from one element
of structure in a text to the next will be signaled in the rst sentence of the new element of
the structure, but not necessarily in the Theme of the rst sentence. This hypothesis seems
to imply that the element of text structure is itself a unit which may have an analog to
thematic structure on the levels of clause and clause complex. Martin (1992: 437 . and 1993
in Halliday and Martin 1993: 244-257) makes clear use of this notion when he discusses
Macrotheme and Hypertheme. I, however, do not know of any empirical study of the use (or
non-use) of Macrotheme and Hypertheme in randomly selected texts.
Themes is part of the signaling. (One possible interpretation of their results

is that the presence of the order sc + mc itself (regardless of the content of
the sc the Theme of the clause complex) was sucient to signal a dis-
course boundary.) As a result, this experiment only suggests that hypothesis
4b is worth exploring.
2.5 Summary
In summary, while previous work has made considerable progress in explor-
ing the four hypotheses, we need considerable expansion of the data which
are used to test them. Even the most robust of the studies mentioned above,
those of Bcklund, Francis, and Plum, use limited data. The other studies are
based on data so restricted as to be of limited value except as support for
trends found in other studies. For example, the two class projects by students
clearly could not stand on their own because of the exceedingly small size
of the samples used. Further, we need considerably more careful description
of factors which can help explain the nature of the texts used as data. Bck-
lunds study suers from the fact that no major eort is made to describe the
contexts in which the conversations are being held in anything more than the
most general terms. Francis attempts to deal with three dierent genres, but
simply uses location within the newspaper as the sole criterion for member-
ship in a genre. Her data show a great deal of variation within each of her
genres, and we are therefore not sure whether her criterion is sucient to
isolate uniform text types. We need much more data taken from a range of
more carefully dened genres and generic structures to use as a basis for our
conclusions. With this aim in view, in the next Section I shall describe some
details of an investigation into the thematic patterning found in certain text
types.
3. Theme in discourse: A further study

As a step toward increasing the breadth of data on which to base our conclu-
sions concerning the relation between the experiential content of Themes and
the organisation of text, and as a means of exploring the complexities of
assigning particular texts to particular genres, I examined the Themes in a
number of short texts which involved narration. I have chosen varieties of
narration so that all texts would be seen to have some comparable features.
340 fries
However, while each text involved narration, each was also produced in
special conditions which imposed special constraints on it.
3.1 The Data

The texts examined fall into four major groups which might be loosely called
super-genre. The rst group obituaries consists of three subgroups
which I will call texts. I analyzed ve obituaries from Time Magazine, two
from the New York Times, and four from The Morning Sun (the local news-
paper in Mount Pleasant, Michigan). The second group consists of a single
text, a program note which described the performer at a concert which I
attended. The third group consists of narratives and includes a fairy tale,
namely Little Red Riding Hood, the rst chapter of A Farewell to Arms by
Ernest Hemingway, and two short historical accounts one an account of
Napoleons invasion of Russia, and the second The Bathtub Navy. Finally,
the fourth group consists of one expository text, called Balloons and Air
Cushion the Fall, which is a report of a new invention. Since I am contrast-
ing the ve obituaries which were found in Time Magazine with other texts
as a group, I will sometimes refer to the Time obituaries as if they constituted

a single text. Similarly, for the same reason, I will often refer to the two
obituaries from the New York Times and the four obituaries found in The
Morning Sun as if each group constituted one text. It is important, however,
to remember that these three texts are actually groups of short texts and so
will exhibit special properties.
We must be careful to notice that while all the texts in the data involve
narration, they do not constitute a single genre, even though they share
certain characteristics. For example we might be tempted to classify all
obituaries as Recounts (Martin and Rothery 1986). But these Recounts are
produced under dierent conditions which aect the language used. At a
very general level, texts written under strong space constraints typically have
a very high lexical density. Thus, the obituaries from Time Magazine use
many grammatical devices (such as appositives) which increase the informa-
tion conveyed per clause, and they contain relatively many clauses per
sentence. The obituaries found in The Morning Sun do not contain many
instances of such condensing devices. By the same token, we might think
that all obituaries have the same structure. However, the ones I examined
were found in three importantly dierent contexts and contained three
dierent structures. The ones taken from Time Magazine were from a section
labeled Milestones. Milestones include more than simply deaths: one can
be Released, Married, Divorced etc. As a result the rst word of each text
indicates which milestone has been passed. All the obituaries taken from
Time Magazine begin with the bolded word died (the Theme of the rst
clause), and then in smaller bold print give the name of the person who has
died. Finally the rst sentences typically concluded by describing the
circumstances under which the death occurred and the location of the death.
By contrast all the other obituaries were taken from sections of the newspa-
pers where the heading of the section made it clear that the entire section
dealt with deaths. In these cases, the Theme of the rst clause indicated who
had died. However, these obituaries also fell into two groups, depending on
their source. The Morning Sun is a local newspaper. The obituaries are of
interest to area inhabitants and so, tend to mention information relevant to
the potential actions of the readers. Thus, one nds information about how
the family can be contacted, where the services will be held, who will
ociate, etc. The people described in these obituaries are usually not
prominent, but merely local residents. These obituaries seem to focus on
when and where things happened; thus we may have comments such as Mr.
Baumgardner was born August 1, 1921 in Macomb il , to. .. and Moving to
Mount Pleasant in 1951, he By contrast, The New York Times, while
serving a local community, is at the same time a national newspaper and is
distributed nationally, and it prints local, national and international news.
One nds only prominent people mentioned in its obituary section. The
actions typically described are ones that have aected many people. The
New York Times obituaries typically do not include references to where and
how the family may be contacted, where the memorial services will be held,
etc.
The program note resembles the obituary from the New York Times in
that it attempts to present aspects of the expertise of the performer. Of course
there is the obvious dierence from an obituary in that the program note
focuses on the present credentials and abilities of the performer. (This focus
is most easily noticed in the verb forms of the text.)
Within the group of texts I have labeled Narratives Little Red Riding
Hood is a fairy tale told to children. As a result, we can infer that the struc-
ture will be relatively simple and straight forward. By contrast, the other
three texts I have labeled narratives are written for adults and we can expect
342 fries
them to be more complex, particularly in their interpersonal stance. Again,

calling these three texts narratives conceals some signicant dierences.
The Napoleon text and the Bathtub Navy are both accounts of historical
events. The Bathtub Navy was published in the Readers Digest as a complete
account. The Napoleon text is a short excerpt from a longer text and more
closely resembles a recount. The Hemingway text is the rst chapter from
a novel. Of course, the entire novel constitutes the narrative. This chapter
merely sets the scene for the narrative. Indeed, it is very dicult to nd any
real events in which a single participant (or small group of participants) did
anything in this chapter. In other words this chapter is not itself what one
would typically call a narrative. Finally, Balloons and Air describes an
invention by a company which solves a problem. This text involves narration
in that it describes a set of events. However it is not a simple narrative. It is
clearly structured in a Situation Problem Solution Evaluation structure.
More to the point, this text clearly focuses on the problem and its solution
rather than on the people who found the solution. (Compare this aspect of
this text with The Bath Tub Navy, which has a similar structure, but which
focuses on the participants involved to a much greater degree.) As a result,
the Balloons and Air narrative is best treated as a type of expository text. As
a nal comment on these texts, I should mention that all of these texts show
strong evidence of having been carefully edited. Indeed the obituaries from
Time and The Morning Sun seem to have been written according to formula.
3.2 Hypothesis 3 revisited

How are these aspects of the context reected in the language of the texts
and particularly in the experiential content of the clause Themes? The search
for ecient ways of expressing what is said which is typical of national
news media is evident in the number of hypotactic clause relations used in
news, and also in the amount of ellipsis particularly the ellipsis of Subjects
in branched clauses. The amount of hypotaxis and Subject ellipsis in a text
can be measured by nding the ratio of non-embedded clauses per t-unit.
The texts found in Time and The New York Times contained 1.9 non-embed-
ded clauses and 1.8 non-embedded clauses per t-unit, respectively. Except
for Little Red Riding Hood, the other texts had signicantly fewer non
embedded clauses per t-unit, ranging from 1.2 to 1.5 clauses per t-unit.
Little Red Riding Hood had 1.8 non-embedded clauses per t-unit a rela-
tively high number. Much of this number came from the 22 quotations and
the 10 branched clauses with Subject ellipsis which occurred in the text.
These features particularly the quotations were rarely found in the other
texts in my data. If these clauses are excluded from the ratio, the text would
show a ratio of 1.24 clauses per t-unit, a ratio that matches the other non-
newspaper narratives.
Let me now move to the grammatical properties of the Themes in the
texts I investigated. The texts in my data made relatively little use of marked
experiential Themes in clauses. Percentages ranged from 2% for the obituar-
ies in The Morning Sun to 36% marked Themes in the clauses of the Napo-
leon extract. None of the texts in my data use embedded clausal Themes,
thematic equatives, or predicated Themes. Few texts in my data used nomin-
alizations as Theme. Three texts use one thematic nominalization each. One
set of texts, the Morning Sun obituaries contains nine nominalizations, but
they are closely associated with the schema which is invoked by the situa-
tion. (Burial will be in Riverside Cemetery, and Memorials may be made
to ) The texts in my data therefore do not encode author intervention
through the thematic use of embedded clause Themes, thematic equatives,
predicated Themes or nominalizations as Theme. These results should be

compared with the gures Francis presents for her data in which she nds
several embedded clause Themes, thematic equatives, predicated Themes
and nominalizations as Themes in her hortatory and analytical exposition
texts, but she nds few of these structures in the news reports she examined.
Again, like the studies summarized in section 1, I found many more
simple Themes than I did multiple Themes. As many as 89% of the clause
Themes from the Morning Sun obituaries were simple. The lowest percent-
age of simple Themes was found in Chapter 1 of Hemingway A Farewell to
Arms, where simple Themes made up only 47% of the total Themes. When
multiple Themes occurred in the texts, they were typically combinations of
textual and experiential Themes. The Hemingway chapter contained the
highest percentage of textual Themes. 49 % of the clauses from the Heming-
way text contained textual Themes. Most of the textual Themes in the texts
consisted of markers of addition, contrast, or of sequence.
Only 11 interpersonal Themes occurred in the data. Five of those 11
occurred in the conversations in Little Red Riding Hood. The remaining 6
were found in the Bathtub Navy, Farewell to Arms, and Balloons and Air.
These texts have relatively complicated interpersonal stances. The compli-
344 fries
Table 6A: Summary of grammatical properties of Themes in various texts involving narration
Genre
Obituaries Program Narratives Expos
Time ny Times Morn Sun Blurb lrrh Napoleon bt Navy Farewell Ballns Total
Clauses 39 44 57 20 98 22 19 49 19 367
Length of Text
t-units 21 24 46 15 53 18 13 40 13 243
Density Cl/t-unit 1.86 1.83 1.24 1.33 1.85 1.22 1.46 1.23 1.46 1.51
Grammatical Simple Themes 20(51%) 34(77%) 51(89%) 17(85%) 68(69%) 14(64%) 13(68%) 23(47%) 14(74%) 254(69%)
properties of
overt Themes Mult. Themes 9(23%) 10(23%) 5(9%) 3(15%) 27(28%) 8(36%) 6(32%) 24(49%) 5(26%) 97(26%)
cl 8(21%) 4(9%) 1(2%) 6(30%) 5(5%) 8(36%) 5(26%) 14(29%) 2(11%) 53(14%)
Marked themes
t-un 4(57%) 1(21%) 1(4%) 2(53%) 2(13%) 0(44%) 2(54%) 1(38%) 1(23%) 14(28%)
Embdded clause Themes 0 0 0 0 0 0 0 0 0 0

Thematic Equatives 0 0 0 0 0 0 0 0 0 0
Predicated Themes 0 0 0 0 0 0 0 0 0 0
Nominalizations as Theme 0(0%) 2(5%) 9(16%) 0(0%) 0(0%) 1(5%) 1(5%) 0(0%) 1(5%) 14(4%)
Interprsonl Theme 0(0%) 0(0%) 0(0%) 0(0%) 5(5%) 0(0%) 2(11%) 2(4%) 2(11%) 11(3%)
On
T Subject
extual Themeand Theme : A discourse
9(23%) functional
8(18%) perspective,
5(9%) edited
3(15%) 25by Ruqaiya
(26%) 10Hasan,
(45%) 5and
(26Peter
%) 24H.
(49%) 4(21%) 93(25%)
addition
Created /sequence 6(15%) 11:02:16.
from coventry on 2017-12-18 7(16%) 4(7%) 1(5%) 22(22%) 9(41%) 2(11%) 22(45%) 2(11%) 75(20%)
(and, but, then)

Table 6B: Summary of semantic properties of Themes in various texts involving narration
Genre
Obituaries Program Narratives Expos
Time ny Times Morn Sun Blurb lrrh Napoleon bt Navy Farewell Ballns Total
Ideational content of Themes:
Retrospective Text References 0 0 0 0 0 0 0 0 1 1
Speaker/Hearer 0 0 0 0 9 0 0 0 0 9
Major Text Participant 16(41%) 13(30%) 39(8%) 7(35%) 30(31%) 13(59%) 7(37%) 1(2%) (0%) 126(34%)
Abstract concept 0(0%) 3(7%) 9(16%) 0(0%) 0(0%) 1(5%) 1(5%) 3(6%) 4(21%) 21(6%)
Process:
a: total 11(28%) 5(11%) 3(5%) 3(15%) 27(28%) 0(0%) 3(16%) 1(2%) 2(11%) 55(15%)
b: non- hypotactic clause 6(15%) 4(9%) 3(5%) 3(15%) 6(6%) 0(0%) 3(16%) 0(0%) 2(11%) 27(7%)
c: verbal process 0 0 0 0 15 0 0 0 0
Temporal location:
a: Clause theme 4(10%) 1(2%) 1(2%) 6(30%) 3(3%) 4(18%) 2(11%) 4(8%) 0 25(7%)
b: t-unit theme 7(33%) 2(8%) 2(4%) 7(47%) 5(9%) 5(28%) 4(31%) 4(10%) 0 36(15%)
Spatial location 0(0%) 1(2%) 0(0%) 0(0%) 2(2%) 1(5%) 3(16%) 6(12%) 1(5%) 14(4%)
On Subject
Subject ellipsesand Theme : A discourse
6(15%)functional
10(23%) perspective,
4(7%) 0edited
(0%) by10Ruqaiya
(10%) Hasan,
2(9%) and
0(0Peter
%) 1H.(2%) 0(0%) 33(9%)
Created from coventry on 2017-12-18
Other 1 11:02:16.
3 0 0 19 0 1 0 1
346 fries
cated interpersonal stances might be expected to increase the use of interper-

sonal elements as Theme. For example, the only interpersonal Theme which
is also a t-unit Theme in the Balloons and Air text is found at a point where
the author evaluates a portion of the situation as a problem.
Tables 6a and 6b summarize the results discussed in this section. The
four major super genres are separated into groups of columns. The individ-
ual texts are given abbreviated names and gures for each text are placed
within individual columns. Table 6a presents grammatical features of
Themes in the data, while Table 6B summarizes semantic features. The
percentages indicate the percent of clauses (or t-units) in a text which exhibit
some feature. For example, the Time obituaries contain 9 textual Themes.
Since these obituaries contain 39 clauses, 23% of the clauses of the Time
obituaries contained a textual Theme. Only overt (i.e. non-elliptical) Themes
have been taken into account, hence percentages comparing the occurrence
of simple and multiple Themes do not always total 100%. The rows labeled
Marked Themes provide the occurrence of marked experiential Themes in
clauses and in t-units. The percentage below the number of marked clause
Themes is the percent of clauses which have marked Themes. The number
given in the row for marked t-unit Themes gives the number of hypotactic
clauses which function as Theme for their t-units. (Hypotactic clauses would
not be considered Themes of clauses.) Since in these data all the marked
clause Themes are also marked t-unit Themes, it is necessary to total the
number of marked clause Themes with the number of marked t-unit Themes
to obtain the percentage of t-units which contain a marked Theme of some
sort. The percentages beside the gures Marked t-unit Themes give the
percentage of t-units which have marked Themes.
3.3 Hypotheses 2 and 3 revisited

We have said earlier that the experiential content of the Themes contributes
to what I have called the method of development of a text. This is the place
where hard counts of data are weakest. As a rst approximation to this issue
Table 6b classies various meanings which were used as experiential
Themes in the texts. Clearly the major text participants (for example refer-
ences to Little Red Riding Hood and to the wolf) accounted for many
experiential Themes in the data. The two major exceptions to this trend,
Farewell to Arms and Balloons and Air, were expected to depart from this
trend since they dier seriously from normal narrative. The Balloons and Air
text did not have an obvious major participant; the text concerned a problem
and how the problem is being solved. While the Hemingway text is from a
narrative which has recognizable major participants, the section of the text
which I analyzed never really got to the story. As mentioned above, the
chapter I examined simply presents the setting for the story. In all the other
texts, references to the major characters account for at least 31% of the
Themes (in Little Red Riding Hood) and accounts for over 40% of the
Themes in four texts (the three groups of obituaries, and Napoleon).
Since all of the texts in my data involve narration, we might expect that
references to temporal and/or spatial locations would be used to move the
texts along. With the exception of the expository text Balloons and Air (the
text we would expect to be least likely to foreground these meanings) all
texts in the data contain at least one thematic reference to temporal or spatial
location. However, the number of references to temporal and spatial location
clearly does not constitute a single method of development in any of these
texts in the sense that I intended this term in my early work. Further, tempo-
ral and spatial location are foregrounded to dierent degrees in the texts and
in dierent ways in the texts. None of the texts in my data even approaches
the thematic use of spatial location that Nie found in his travel guide for
Hangzhou, and the travel brochures which Margaret Berry studied also used
spatial location signicantly more as Themes of clauses than any of my
texts. Table 7 summarizes every non-embedded reference to temporal or
spatial location which occurred in the clauses of my data. (For ease of
comparison I have included at the bottom of Table 7 the comparable gures
for Berrys study and Nies study. ) A non-embedded reference is one which
is not a part of a larger group or phrase. The bold portions of Examples 1-4
below are non-embedded references to temporal or spatial location, while
the italicized portions of the examples are embedded references to temporal
and spatial locations and were not counted. These are reproduced below:
(i) Mr. Brand, who was born in Kewanee il, joined the Army and served as a
platoon sergeant in World War ii,
(ii) Moving to Hollywood, he made his lm debut in 1950 as a hoodlum
(iii) His rst attempt at acting was in a 1946 Army Signal Corps movie
(iv) As the morning sun lighted the beaches of Dunkirk, the rst of hundreds of
small boats pulled onto the shore.
348 fries
Table 7: References to spatial and temporal location
Theme Other N-Rheme Total
Meaning Choice: Location Temp Spa =Role Temp Spa =Role Temp Spa =Role Temp Spa =Role
Genre type & source
Time 2 0 - 1 4 - 6 7 - 9 11 -
Obituary ny Times 0 0 3 2 0 1 11 8 3 13 8 7
Morn Sun 1 0 - 16 7 - 8 25 - 25 32 -
Program Note 4 - - 1 1 - 1 9 - 6 10 -
lrr Hood 3 2 - 0 2 - 4 23 - 7 27 -
Napoleon 3 1 - 0 0 - 0 7 - 3 8 -
Stories
bt Navy 3 3 - 0 1 - 0 6 - 3 10 -
Hemingway 4 4 - 2 3 - 1 13 - 7 20 -
Exposition Balloons & Air 0 1 - 1 0 - 1 1 - 2 2 -
25
Coee Party 2 25
Committee Meeting 3 25
Berry Texts
Travel Brochure 10 25
Guide
Fries, John book
Benjamins 14
Publishing Company, 25
1995. ProQuest Ebook Central, http://ebookcentral.proquest.com/lib/coventry/detail.action?docID=
Nie Long Text Hzo Guide book 1 6 0 3 9 5 10 14
I divided each clause in my data into three sections: Theme, n-Rheme,

and Other. The Theme of a clause is the initial constituent of that clause and
provides a framework for the interpretation of the message expressed by the
clause. The n-Rheme of a clause is the nal constituent of the clause, and
is of interest because it is the location of the unmarked placement of New
information. Since clauses usually contain more than two constituents (and
so the combination of Theme and n-Rheme typically does not always
exhaust the clause), I added an Other category for material which was
neither Theme nor n-Rheme.
The Morning Sun obituaries and Little Red Riding Hood resemble
Berrys Coee Party and Committee meeting texts in that none of these texts
uses Spatial Location very often within the clause Themes. This is not to say
that these texts do not refer to Spatial Locations, rather the spatial references
typically form part of the focus of the messages and therefore occur regularly
in the n-Rhemes of the clauses. The Hemingway text, with its focus on
description of a scene, vaguely resembles a guidebook in its orientation to
spatial location and this fact is reected in its somewhat greater use of
Spatial Location as Theme. However, Hemingways writing is still very far
from the guidebook genre which Berry and Nie found.

The texts I examined contained a number of thematic references to
processes (up to 28 percent of the clauses of some texts referred to processes
in the Themes.) The major source of these thematic references to process
consisted of non-nite clauses beginning with verbs (e.g. the bolded clause
in Mr. Baumgardner worked for Federal Mogul for 35 years, retiring in
1978.). Of course, the imperatives in Little Red Riding Hood also added to
the number of verb initial clauses, and, nally, Little Red Riding Hood
contained 15 instances of quoting verbs which began their clauses. (e.g.
And where does your grandmother live? asked the wolf politely.) This
ordering of verbal process and Sayer seems strongly associated with fairy
tale language14. It is useful to note, however, that this inversion of Verbal
process and Sayer was not found at the beginnings of the stories I have
looked at. Rather, it began once the story was under way. In the case of Little
Red Riding Hood, once the story was begun and the rst Subject Predicator
14. Of course inversion of Subject and predicator in quotatives is not limited to fairy tales.
News reports often contain instances of this sort of inversion.
350 fries
inversion was found, 15 of 18 quotative clauses used the Verbal Process as

Theme15.
3.4 Hypothesis 1 revisited

Finally, let us examine the evidence for the relation between thematic
progression and genre type. One factor involved in thematic progression
concerns the chain relations in which the various Themes of a text partici-
pate. As a means of exploring this issue I examined nominal groups which
were found as part of the Themes of the clauses of the texts in my data and
asked whether these nominal groups participated in chains. If the answer was
yes, I then located the most recent previous member of the chain. Speci-
cally, was the most recent previous member of the chain located in the
Theme, Other, or n-Rheme portion of the earlier clause? Text 1 provides
some examples of how I counted the data.
Text 1: A Farewell to Arms
1. In the late summer of that year we lived in a house in a village that looked
across the river and the plain to the mountains.

2a. In the bed of the river there were pebbles and boulders, dry and white in the
sun,
2b. and the water was clear and swiftly moving and blue in the channels.
3a. Troops went by the house and down the road,
3b. and the dust they raised powdered the leaves of the trees.
4a. The trunks of the trees too were dusty
4b. and the leaves fell early that year
4c. and we saw the troops marching along the road
5a. The plain was rich with crops,
Each punctuated sentence in Text 1 is given a number and begins a new
line. If a sentence contains more than one non-embedded clause, then each
15. In all cases of inversion of verbal process and Sayer, at least part of the quotation
preceded the quoting clause. As a result, the general nature of the process in the quotative
was not at issue. It was clear from the context that it would be some sort of verbal process.
What was at issue was the specic identity of the Sayer. In other words, one important factor
in the use of these inversions seems to be the placement of Sayer in a position of focus
relative to the process. (A further factor relevant to the analysis of these constructions is that
the quotatives would probably be read as a minor tone group (Tone 3 in a 1-3 combination).
non-embedded clause is given a letter and begins a new line.

Obvious members of chains are the leaves (Theme of 4b) which forms
part of an identity chain which has the leaves of the trees in the n-Rheme
of 3b as its last previous occurrence, and the plain (Theme of 5a) which
forms part of an identity chain which has the plain in the n-Rheme of 1 as
its last previous occurrence. Slightly less obvious is the bed of the river
(Theme of 2a) which forms part of a similarity chain which has the river in
the n-Rheme of 1 as its last previous occurrence. The water (the Theme of
2b) is the next member of the same similarity chain. In the case of 2b (the
water), the last previous member of the similarity chain is in the Theme of
2a. The Theme of 4a, (the trunks of the trees) also belongs to a similarity
chain. The last previous mention of the chain is found in the n-Rheme of 3b
(the leaves of the trees). If we now focus on the coding of the data, the
Theme of 2a (in the bed of the river) contains an item which belongs to a
chain which has as its last previous member an item which has been men-
tioned in the n-Rheme of a clause in the previous sentence. It was therefore
coded as n-Rheme of the previous sentence (= Prev s). The Theme of 5a
(the plain) belongs to a chain whose last member occurred in the n-Rheme
of the clause in sentence 1. Three sentences intervene between Sentence 1

and Sentence 5, hence this item was coded as n-Rheme, Skip 3. Some items
belonged both to similarity chains (e.g. the leaves in 4b can be seen as part
of a similarity chain with the trunks of the trees in 4a) and to identity chains
(the leaves in 4b can be seen as part of an identity chain with the leaves of
the trees in 3b). In these cases, the identity chain relation was counted, not
the similarity chain relation.
Tables 8a, b, and c summarize the results of the counts of thematic
progression by major text group.
Clearly the obituaries display a strong tendency to use a Theme iteration
(Theme Theme) progression. Exceptions to this pattern are concentrated
in the Time Magazine obituaries. These exceptions result from the require-
ment that every obituary in Time Magazine begin with the Theme Died.
Because of this requirement, all the remainder of the rst sentence is either
n-Rheme or Other. Since the second sentence of each of these obituaries
elaborates on the person who died (not on the nature of the death), the
second sentence in every obituary exemplies either an n-Rheme Theme
progression or an Other Theme progression (two variants of Linear
thematic progression). The eect of this formula for writing obituaries in
352 fries
Table 8a: Last Previous mention of Nominal Themes: Obituaries

Same s Prev s Skip 1 Skip 2 Skip 3 Skip >3 Total
Prev Cl
Theme
Morning Sun 30 1 4 3 1 39
Time 2 8 10
N.Y. Times 1 9 10
Total 3 47 1 4 3 1 59
Other
Morning Sun 0
Time 3 3
N.Y. Times 2 1 3
Total 5 1 6
n-Rheme
Morning Sun 0
Time 1 3 4
N.Y. Times 1 1
Total 1 4 5
Time is increased by the fact that all the Time Magazine obituaries are very
short. The longest Time obituary in my data contains only six sentences.
Most contain only three sentences. A further consequence of the brevity of
the Time obituaries is that it was rarely possible to have chains which
skipped more than two sentences in these texts.
Little Red Riding Hood (lrrh) shows a similar pattern of theme ite-
ration progression, though 14 (a signicant minority) of the clause Themes
in that story relate most recently to the n-Rheme of some previous clause.
Three of these 14 n-Rheme Theme sequences result when the most recent
previous mention of a Theme is found in a quotative segment which has
undergone Subject Predicate inversion, and three others are found when
the most recent mention of a Theme is found in a vocative at the end of a
previous quotation. Both of these situations constitute special factors which
aect this text but not the others in my data. The Napoleon text and The Bath
Tub Navy (bt Navy) also show a predominance of Theme iteration progres-
sions, however the Hemingway chapter exhibits a strong trend towards a
linear (n-Rheme Theme) progression. However, I have already pointed
out that this chapter is not truly a narrative, but rather the setting of the scene
for a narrative.
In summary, then, the narratives reported on here show a general
progression from a Theme iteration (Theme Theme) progression in the
story written for children toward a greater use of a linear (n-Rheme
Theme) progression in stories written for adults. The dierence in Thematic
progression results from the fact that the stories written for adults achieve
a variety of tasks other than simple narration. If one looks at Text 1 taken
from the Hemingway chapter we see that signicant portions of this chapter
can be viewed as elaborations of several basic sentences. For example,
clauses 2a, 2b and 5a in Text 1 above can be seen to elaborate parts of clause
1, and the Themes of these three clauses direct the readers attention back to
the points of elaboration. The points being elaborated are located in the n-
Rheme of clause 1. Roughly similar factors contribute to the use of
Table 8b: Last previous mention of nominal Themes: Stories

Same S Prev S Skip 1 Skip 2 Skip 3 Skip >3 Total
Prev Cl
Theme
lrrh 3 16 7 1 27
Napoleon 4 3 7
bt Navy 1 2 2 1 2 8
Farewell 7 4 3 14
Other
lrrh 2 1 3
Napoleon 2 1 1 4
bt Navy 2 2
Farewell 1 1
n-Rheme
lrrh 2 7 3 1 1 14
Napoleon 3 1 4
bt Navy 1 1 2
Farewell 3 7 1
3 14
354 fries
Table 8c: last previous mention of nominal Themes: program

note and Balloons and Air
Same S Prev S Skip 1 Skip 2 Skip 3 Skip >3 Total
Prev Cl
Theme
Prog. Note 1 4 5
Balloons 1 1 1 3
Other
Prog. Note 2 2
Balloons 1 1
n-Rhem e
Prog. Note 2 2
Balloons 1 2 1 4
n-Rheme Theme and Other Theme progressions in the other narrative

texts. Thus, three of the seven instances of these patterns in the Napoleon text
occur in descriptions of the situation which obtained at the time being

described.
Finally, Balloons and Air, the sole example of expository writing, shows
a tendency to use linear (n-Rheme Theme) progressions. In three out of
the four instances of this pattern, the second member of the pair elaborates
a portion of the rst member, and the cohesive tie in which the Theme
participates indicates what is being elaborated.
4. Conclusion
This paper has explored hypotheses 1 and 3. A serious examination of
hypothesis 2 would require an investigation of reader/listener responses to
texts, which is beyond the scope of this study. Similarly, hypothesis 4, the
relation of the experiential content of the Theme to the generic structure of
texts, has not been systematically explored. The data discussed in this paper
and in the works summarized in section 1 make it quite clear that hypotheses
1 and 3 are worth exploring carefully. Thematic progressions and the
experiential content of the Themes do not occur randomly in these texts. In
my data, the frequencies of the various thematic progressions vary with
genre type, the experiential content of the Themes varies with genre type,
and the proportions of times that certain meanings are expressed thematically
also varies with genre type. Further, while my data did not support a careful
investigation of the relation of these features to the generic structure of
texts16, the data did suggest that such a study would be protable17. Having
said that the thematic progressions and the experiential content of Themes
in a text do not occur randomly, the problem remains of describing the
patterns of occurrence in a positive way. These patterns are complicated.
Many of the complications derive from the particular purposes of the text at
a given point. It is not reasonable that we should expect an entire text to do
only one thing. Stories not only must recount events, they must evaluate,
they often tell what could have occurred but did not, and so forth. As the
author changes task and responds to dierent pressures, the Themes of the
clauses of the text will reect and encode those changes. The texts I have
chosen are all narrative in some sense and so may be assumed to encode
some basic similarity of purpose. I have taken pains in the choice of my data,
however, to demonstrate that what may appear at rst glance to be instances
of similar purposes may involve important dierences in actual situation to
which the participants respond. These dierences are encoded in the lan-
guage of the texts. Thus, while Farewell to Arms is a narrative, that portion
of the novel which I analyzed will be seen as a very strange narrative unless
we notice that at that point it is not a narrative, it sets the scene. We should
expect this fact to aect the Thematic choices made in the text at this place.
Because of the nature of the texts I used as data, I have focused on the
presence of elaboration as a source of deviation from ordinary patterns of
thematic progression. It is easy to show that the presence or absence of
elaboration correlates with the contextual construct in my data the obituar-
ies and Little Red Riding Hood (a story which is told to children) have little
elaboration while the other texts use elaboration more often. It is also easy
16. In the original conception of this paper, I wished to use a categorization of verb types such
as the one proposed by Longacre (1983 and 1989) as a means of indicating the dierent
functions which clauses might play in a text. It seemed clear to me that such a categorization
would correlate with dierent thematic content. It was not possible to explore these issues
in the present paper, however, because I did not choose data which exhibited sucient
variation in internal structure or in information types to make it worth exploring.
17. See Cloran this volume for some interesting suggestions.

356 fries
to see that the presence of elaboration correlates with a certain type of

experiential content of the clause Themes and thematic progression. Thus,
in the data discussed in section 2 of this paper, many of the exceptions to the
general patterns of thematic progression and the patterns in the experiential
content of the Themes are related to genre types.
References
Bcklund, Ingegerd. 1989. To sum up: Initial innitives as cues to the reader. Paper
presented at the Sixteenth International Systemic Congress. June 12-16, 1989.
Helsinki, Finland.
Bcklund, Ingegerd. 1990. Theme in English telephone conversation. Paper de-
livered at the 17th International Systemic Congress, Stirling, Scotland. June
1990.
Benson, James D., William S. Greaves, and Glen Stillar. 1992. Motivating text and
construing context in two lm reviews. Paper presented to the 19th International
Systemic Functional Congress, July 13-18, 1992.
Berry, Margaret. 1987. The functions of place names. Leeds Studies in English, New
Series xviii: Studies in honour of Kenneth Cameron, edited by Thorlac Tur-
ville-Petre and Margaret Gelling, 71-88. Leeds: School of English, University

of Leeds.
Brown, Gillian, and George Yule. 1983. Discourse Analysis. Cambridge: Cambridge
University press.
Chafe, Wallace. 1976. Givenness, contrastiveness, deniteness, subjects, topics, and
point of view. Subject and Topic, edited by Charles Li, 27-55. New York:
Academic Press.
Chafe, W.L. 1984. How people use adverbial clauses. Proceedings of the tenth
annual meeting of the Berkeley Linguistics Society Berkeley: Berkeley Linguis-
tics Society, 437-449.
Dane, Frantisek. 1974. Functional sentence perspective and the organisation of the
text. Papers on Functional Sentence Perspective, edited by Frantisek Danes,
106-128. The Hague: Mouton.
Downing, Angela. 1991. An alternative approach to Theme: A systemic functional
perspective. Word 42.2, 119-143.
Francis, Gill. 1989. Thematic selection and distribution in written discourse. Word
40, 201-221.
Francis, Gill. 1990. Theme in the daily press. Occasional Papers in Systemic
Linguistics 4, 51-87.
Fries, Peter H. 1981. On the status of Theme: Arguments from discourse. Forum
Linguisticum 6(1), 1-38.
Fries, Peter H. 1992. The structuring of written English text. Current Research in
Functional Grammar, Discourse, and Computational Linguistics with a Foun-
dation in Systemic Theory, edited by M.A.K. Halliday and F.C.C. Peng. Special
Issue of Language Sciences 14.4, 1-28.
Fries. Peter H. 1993. Information ow in written advertising. Language, Communi-
cation and Social Meaning, edited by James Alatis, 336-352. Georgetown
University Round Table on Languages and Linguistics, 1992. Washington dc:
Georgetown University Press.
Fries, Peter H. in press a. On Theme, Rheme and discourse goals. Advances in
Written Text Analysis, edited by Malcolm Coulthard, 229-249. London: Rout-
ledge and Kegan Paul.
Fries, Peter H. in press b. Patterns of information in initial position in English.
Discourse and Meaning in Society: Functional perspectives edited by Peter H.
Fries and Michael Gregory. Norwood nj: Ablex Publishers.
Gundel, Jeanette. 1977. The role of topic and comment in linguistic theory.
Bloomington, in: Indiana Linguistics Club.
Halliday, M.A.K. 1967. Notes on transitivity and theme in English: Part 2. Journal
of Linguistics 3.2, 199-244.
Halliday, M.A.K. 1970. Language structure and language function. New Horizons
in Linguistics, edited by John Lyons, 140-164. Harmondsworth: Penguin.

Halliday, M.A.K. 1985. An Introduction to Functional Grammar. London: Edward
Arnold.
Halliday, M.A.K. 1989. Spoken and Written Language. London: Oxford University
Press.
Halliday, M.A.K. 1993. The construction of knowledge and value in the grammar
of scientic discourse: Charles Darwins The Origin of Species. M.A.K.
Halliday and J. R. Martin, 1993, 86-105.
Halliday, M.A.K. and James R. Martin, 1993. Writing Science: Literacy and
discursive power. London: Falmer Press.
Hasan, Ruqaiya. 1984. Coherence and cohesive harmony. Understanding Reading
Comprehension, edited by James Flood, 181-219. Newark de: International
Reading Association.
Hasan, Ruqiaya. 1985. The texture of a text. In Language, Context and Text: Aspects
of language in a siocial semiotic perspective, by M.A.K Halliday and Ruqaiya
Hasan, 76-96. Geelong, Victoria: Deakin University Press.
Hasan, Ruqaiya. in press. The conception of context in texts. Discourse in Society:
Functional perspectives, edited by Peter H. Fries and Michael Gregory. Nor-
wood, nj: Ablex Publishers.
Huddleston, Rodney. 1988. Constituency, multifunctionality and grammaticalization
in Hallidays functional grammar. Journal of Linguistics 24, 137-174.
358 fries
Huddleston, Rodney. 1991. Further remarks on Hallidays functional grammar: A

reply to Matthiessen and Martin. Occasional Papers in Systemic Linguistics 5,
75-129.
Huddleston, Rodney. 1992. On Hallidays functional grammar: A reply to Martin
and to Martin and Matthiessen. Occasional Papers in Systemic Linguistics 6,
197-212.
Hudson, Richard. 1986. Systemic grammar. Review article. Linguistics 24, 791-815.
Hunt, Kellogg. 1965. Grammatical Structures Written at Three Grade Levels.
Champaign, il: ncte.
Longacre, Robert E. 1981. A spectrum and prole approach to discourse analysis.
Text 1.4, 337-359.
Longacre, Robert. 1983. Vertical threads of cohesion in discourse. Papers in
Textlinguistics 38, 99-111.
Longacre, Robert. 1989. Two hypotheses regarding text generation and analysis.
Discourse Processes 12, 413-460.
Longacre, Robert E. 1990. Storyline Concerns and Word Order Typology. Los
Angeles, ca: Department of Linguistics, University of California at Los
Angeles. (Introduction)
Martin, James. R. 1986. Grammaticalising ecology: the politics of baby seals and
kangaroos. Language, Semiotics, Ideology, edited by T. Threadgold, E.A.

Grosz, G. Kress & M.A.K. Halliday, 225-268. Sydney: Sydney Association for
Studies in Society and Culture (Sydney Studies in Society and Culture 3).
Martin, James R. 1989. Factual Writing: Exploring and challenging social reality.
London: Oxford University Press. Second edition.
Martin, James. 1992. English Text: System and structure. Amsterdam: John Benja-
mins
Martin, James R. and Joan Rothery. 1986. What a functional approach to the writing
task can show teachers about good writing. Functional Approaches to Writing:
Research perspectives, edited by Barbara Couture, 241-265. Norwood, nj:
Ablex.
Mathesius, V. 1939. O tak zvanm aktulnm cleneni vetnm [On the so-called
Functional Sentence Perspective]. Slovo a Slovestnost 5, 171-174.
Nie, Long. 1991. An analysis of texts of dierent genres in terms of thematic
selections and process types. Unpublished Manuscript.
Plum, Gunther. 1988. Textual and contextual conditioning in spoken English: A
genre-based approach. Ph. D. Dissertation. Sydney: University of Sydney.
Prideaux, Gary D. and John T. Hogan. 1993. Markedness as a discourse manage-
ment device. Word 344, 397-411.
Ragan, Peter. 1987. Meaning in the communication of a set of instructions. Ph.D.
Dissertation, National University of Singapore.
Rothery, J. 1990. Story Writing in Primary School: assessing narrative type genres.
Ph.D. Thesis. Department of Linguistics, University of Sydney.
Thompson, Sandra A. 1985. Initial vs nal purpose clauses in written English
discourse. Text 5 1-2, 55-84.
Ure, Jean. 1971. Lexical density and register dierentiation. Applications of Linguis-
tics edited by G.E. Perrin and J.L.M. Trim, 443-452. Cambridge: Cambridge
University Press.
Vande Kopple, W.J. 1991 Themes, thematic progressions, and some implications
for understanding discourse. Written Communication 8, 311-347.
Wang, Ling. 1991. Analysis of thematic variations in Buried Child Paper delivered
to the First Biennial Conference on Discourse. Hangzhou Peoples Republic of
China. June 7-9, 1991.
Wang, Ling. 1992. The theory of Theme and its application to drama analysis. ma
Thesis, Department of English Language and Literature, Peking University.
Winter, Eugene O. 1977. Replacement as a Function of Repetition: A study of some
of its principal features in the clause relations of contemporary English. #77-
70,036, Ann Arbor mi: University Microlms.
Xiao, Qun. 1991. Toward thematic selection in dierent genre: Fables and recipies.
Unpublished manuscript.

On Subject and Theme A Discourse Functional Perspe... - (9. Themes, Methods of Development, and Texts)

Uploaded by

Copyright:

Available Formats

On Subject and Theme A Discourse Functional Perspe... - (9. Themes, Methods of Development, and Texts)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

On Subject and Theme A Discourse Functional Perspe... - (9. Themes, Methods of Development, and Texts)

Uploaded by

Copyright:

Available Formats

What is the paper exploring?

What is the paper exploring?

What are some of the criticisms of the concept of Theme?

What are some of the criticisms of the concept of Theme?

9

Themes, Methods of Development, and Texts*

2. Review of previous work

3. See Bcklund (1989) and Winter (1977:475) for similar wordings.

be the method of development of a text or text segment 4.

Figure 1: Three patterns of Thematic progression

Bcklunds (1990) discussion of eight telephone conversations from the

as Theme (e.g. as described in Francis) might signal some aspect of text

still expect a relatively larger number of references to spatial location in the

processes (the participants), or to the processes themselves were more

Table 1: Distribution of typical lexis of Theme in Francis data

Information type News Editorials Letters

Table 2: Distribution of clausal Themes in Francis data

Of course the structures just listed may be regarded as simply grammatical

content of the clause Themes of hortatory and analytical exposition. In

script of a committee meeting, Berry (1987) provided evidence which

Table 3: Occurrences of place names in Subject,

Table 4: Occurrences of place-names in Theme

Clearly spatial location constitutes an important point of departure for

as Theme. Table 5b provides information on the use of dierent types of

themes, development and texts 333

Marked Themes 6 (21%) 4 (5%) 19 (5%) 0 6 (16%) 5 (13%) 8 (57%)

Major Text 107 (53%) 32 (22%) 52 (28%) 6 (75%) 5 (13%) 26 (67%)

Abstract concept 43 (21%) 61 (41%) 62 (34%) 5 (13%)

Time 10 7 10 1 (13%) 3 (7.7%) 1 (7%)

Hypothesis 4a can in fact be made more specic. As we saw above, the

is the unmarked location of New information, and made the following

in the preparation, and 1 Theme expressed a Circumstance of Manner for the

37 instances of the order mc + sc in the oral narratives occurred after the

Themes is part of the signaling. (One possible interpretation of their results

3. Theme in discourse: A further study

3.1 The Data

as a group, I will sometimes refer to the Time obituaries as if they constituted

them to be more complex, particularly in their interpersonal stance. Again,

3.2 Hypothesis 3 revisited

predicated Themes or nominalizations as Theme. These results should be

Embdded clause Themes 0 0 0 0 0 0 0 0 0 0

themes, development and texts 345

cated interpersonal stances might be expected to increase the use of interper-

3.3 Hypotheses 2 and 3 revisited

Exposition Balloons & Air 0 1 - 1 0 - 1 1 - 2 2 -

I divided each clause in my data into three sections: Theme, n-Rheme,

from the guidebook genre which Berry and Nie found.

inversion was found, 15 of 18 quotative clauses used the Verbal Process as

3.4 Hypothesis 1 revisited

across the river and the plain to the mountains.

non-embedded clause is given a letter and begins a new line.

of the clause in sentence 1. Three sentences intervene between Sentence 1

Table 8a: Last Previous mention of Nominal Themes: Obituaries

Table 8b: Last previous mention of nominal Themes: Stories

Table 8c: last previous mention of nominal Themes: program

n-Rheme Theme and Other Theme progressions in the other narrative

occur in descriptions of the situation which obtained at the time being

17. See Cloran this volume for some interesting suggestions.

to see that the presence of elaboration correlates with a certain type of