Professional Documents
Culture Documents
Qingshun He
Bingjun Yang
Absolute Clauses
in English from the
Systemic Functional
Perspective
A Corpus-Based Study
The M.A.K. Halliday Library Functional
Linguistics Series
Series editors
Chenguang Chang
Guowen Huang
About the Series
13
Qingshun He Bingjun Yang
Faculty of English Language and Culture School of Foreign Languages
Guangdong University of Foreign Studies Shanghai Jiao Tong University
Guangzhou Shanghai
China China
This research could not have been completed without the help of many col-
leagues and friends, among whom we are particularly grateful to Profs. Guowen
Huang and Chenguang Chang at Sun Yat-sen University. Professors Kaibao
Hu (Shanghai Jiao Tong University) and Binli Wen (Guangdong University of
Foreign Studies) generously supported us in many ways during the writing and
revising of the manuscript. Many thanks also to Rebecca Zhu, Yi Xu and Evelyn
Ebina J. from Springer for their help. The research is supported by the Publication
Fund of Guangdong University of Foreign Studies. This research is also supported
by the Program for New Century Excellent Talents in University from China
Ministry of Education (NCET-11-0704) and the Fundamental Research Funds for
the Central Universities (SWU1409102).
Bingjun Yang
v
Contents
1 Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Research Background. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Purpose of the Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Organization of the Book. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
vii
viii Contents
4 Research Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
4.1 Research Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
4.2 Research Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
4.3 Data Collection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
4.4 Data Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
7 Discussions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
7.1 Function. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
7.2 Case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
7.3 Style. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
7.4 Time. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
7.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157
8 Conclusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
8.1 Main Findings of This Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
8.2 Limitations and Further Research . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
8.2.1 Limitations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
8.2.2 Further Research. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
xi
Notational Conventions
xiii
Figures
xv
xvi Figures
xix
xx Tables
Table 6.15 Distribution of absolute clauses in the seven styles in BNC. . . . 124
Table 6.16 Stylistic distributions of with and absolute clauses of extension. . . 127
Table 6.17 Stylistic distributions of subordinators and absolute
clauses of enhancement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
Table 6.18 Stylistic distributions of complementizer that
and absolute clauses of projection. . . . . . . . . . . . . . . . . . . . . . . . 127
Table 6.19 Stylistic distributions of with and absolute clauses
of extension (per hundred million words) . . . . . . . . . . . . . . . . . . 128
Table 6.20 Stylistic distributions of subordinators and absolute
clauses of enhancement (per hundred million words). . . . . . . . . 128
Table 6.21 Stylistic distributions of complementizer that and absolute
clauses of projection (per hundred million words) . . . . . . . . . . . 128
Table 6.22 Stylistic distributions of linkers and absolute clauses. . . . . . . . . 132
Table 6.23 Historical distribution of the absolute clauses in COHA. . . . . . . 133
Table 6.24 Functional distribution of absolute clauses . . . . . . . . . . . . . . . . . 134
Table 6.25 Case distribution of personal pronoun subjects of absolute
clauses in COHA. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
Table 6.26 Functional distribution of the case of absolute clauses. . . . . . . . 139
Table 6.27 Functional distribution of the case of absolute clauses over time. . . 140
Table 7.1 Absolute clauses used independently. . . . . . . . . . . . . . . . . . . . . . 144
Table 7.2 Absolute clauses used independently in COHA . . . . . . . . . . . . . 145
Table 7.3 Case-marked personal pronoun subjects. . . . . . . . . . . . . . . . . . . 148
Table 7.4 Absolute clauses of enhancement in three corpora . . . . . . . . . . . 154
Table 7.5 Most frequent absolute clauses of enhancement in COHA. . . . . 155
Chapter 1
Introduction
1.1 Research Background
However, absolute nominative clauses in late Middle and Modern English were
nominative case. “About the middle of the fourteenth century, the nominative
began to replace the dative.” (Morris 1886[2010]: 103), while the dative “lasted
down to 1400” (Oliphant 1878: 408). In fact, Wycliffe ever used nominative pro-
nouns in his translation, indicating that the case of absolute nominative clauses
had begun to change into nominative before Wycliffe, which was probably com-
pleted in 1420s or 1430s (Poutsma 1929: 973). Modern English grammarians gen-
erally hold that the subjects of absolute nominative clauses are zero case nouns
or nominative pronouns, sometimes accusative pronouns (Curme 1931: 154;
Jespersen 1949: 45; Visser 1972: 1148), but very few in number, some of which
are obviously affected by Latin. Some other grammarians think that the accusative
form is often considered unacceptable in Modern English (Fowler 1965: 4; Stump
1985: 11). Such grammarians as Quirk et al. (1985) and Biber et al. (1999) use
“absolute clauses” to name absolute nominative clauses to avoid the case prob-
lem. To some extent, they accept the fact that the subjects of absolute nominative
clauses can be accusative.
Absolute nominative clauses began to appear in English poems in the four-
teenth century and became widely used in the seventeenth and eighteenth centu-
ries, but were gradually reduced in the nineteenth century and then confined to
formal styles (Ross 1893: 14–15). As to style, it is generally believed that abso-
lute nominative clauses “is not only uncolloquial, but is by many felt to be un-
English, and to be avoided in writing as well” (Sweet 1903: 124). “Except for
stereotyped phrases like weather permitting, the colloquial use of the Nominative
Absolute is almost restricted to it being…, there being…” (Onions 1905[2010]:
67). Grammarians nowadays seldom carry out specialized research on absolute
nominative clauses and only discuss them in grammar books as a certain gram-
matical structure. However, absolute nominative clause is a very common gram-
matical structure (Wooley 1920; Pence and Emery 1965) and absolute nominative
clauses are the most commonly used sentence modifiers (Al-Hamash and Abdulla
1979: 379) in actual English usages.
Traditional grammarians are also interested in the function types of absolute
nominative clauses. Curme (1931: 154–157) distinguishes six broad logical roles
according to whether an absolute is intuitively linked to its superordinate clause,
including the relations of time, cause, condition and exception, attendant circum-
stance, manner proper, and concession. Kruisinga (1932: 274–275) assumes the
whole range of logical roles under the notion of attendant circumstances. Jespersen
(1949: 61–64) suggests that absolutes can be seen to play four sorts of logical
roles: cause or reason, condition, time, and descriptive circumstances, admitting “it
is often difficult or even impossible to draw sharp boundaries between the several
applications.” Visser (1972) categorizes four logical roles played by absolute nomi-
native clauses: (1) attendant circumstances, (2) reason, ground, cause, or motive,
(3) time, and (4) condition. Quirk et al. (1973: 762) distinguish such three types
of logical relations in absolute clauses as cause, time, and circumstance. Although
grammarians distinguish different function types of absolute nominative clauses,
they all categorize absolute nominative clauses into adverbial clauses.
1.1 Research Background 3
relationship absolute clauses can realize, and the synchronic and diachronic dis-
tributions of absolute clauses in actual language use. We wish that this could not
only make up for the deficiencies of traditional form-based grammatical research
on absolute clauses, but also make some contribution to SFL.
This research consists of three parts. The first part includes the introduction, the
literature review, the theoretical basis, and the research design. The second part is
the SFL research of absolute clauses. The third part is a corpus-based quantitative
research of absolute clauses and the conclusion.
The first part includes four chapters. This chapter is the introduction. Chapter 2
reviews the status quo of the research on absolute clauses in order to reveal the
problems that lie behind the current research. Chapter 3 puts forward the SFL
approach to absolute clauses, offering a sketch of the functional syntactic theory,
and proposes the identification criteria of absolute clauses through a SFL defini-
tion of absolute clauses and a discussion of the conditions of formation. Chapter 4
introduces the research questions, methodology, and the method of data collection
and analysis.
The second part is the fifth chapter. This chapter assumes the meaning potential
of absolute clauses according to the identification criteria proposed in Chap. 3 and
discusses the relationships realized by absolute clauses in the network of clause
complex. Different types of absolute clauses are also different in the tendency to
be independent from the primary clause. The independence of absolute clauses
is embodied on two dimensions, i.e., that of primary clauses and that of absolute
clauses.
The third part includes Chaps. 6 and 7, which is a corpus-based quantitative
study of absolute clauses. Chapter 6 is mainly a research on the relations of elabo-
ration, extension, and enhancement of expansion and that of projection realized by
absolute clauses, involving such aspects as functional distribution, stylistic distri-
bution, historical distribution, and case choice. Chapter 7 analyzes and discusses
the research results in Chap. 6.
Chapter 8 is the conclusion of this research. It summarizes this research, points
out its problems, and suggests future researches in this field.
References
Al-Hamash, K. I., & Abdulla, J. J. (1979). A course in modern English grammar. Baghdad:
IDELTI.
Biber, D., Johansson, S., Leech, G., Conrad, S., & Finegan, E. (1999). Longman grammar of spo-
ken and written English. London: Longman.
Collins, P. C. (1991). Cleft and pseudo-cleft constructions in English. London: Routledge.
References 5
Crystal, D. (2008). A dictionary of linguistics and phonetics (6th ed.). Oxford: Blackwell
Publishing.
Curme, G. O. (1931). A grammar of the English language: Syntax. Boston: D.C. Heath.
Fawcett, R. P. (2000). Theory of syntax for systemic functional linguistics. Philadelphia,
Amsterdam: John Benjamins.
Fawcett, R. P. (2008). Invitation to systemic functional linguistics through the Cardiff gram-
mar: An extension and simplification of Halliday’s systemic functional grammar (3rd ed.).
London: Equinox.
Fawcett, R. P. (2009). Functional syntax handbook: Analyzing English at the level of form.
London: Equinox.
Fowler, H. W. (1965). A dictionary of modern English usage (2nd ed.). Oxford: Oxford
University Press.
Halliday, M. A. K. (1994). An introduction to functional grammar (2nd ed.). London: Edward
Arnold.
Halliday, M. A. K., & Matthiessen, C. M. I. M. (1999). Construing experience through meaning:
A language-based approach to cognition. London, New York: Cassell.
He, W., & Gao, S.-W. (2011). Studies of functional syntax. Beijing: Foreign Language Teaching
and Research Press.
Huang, G.-W. (1999). A functional approach to English syntactic analysis. Journal of Sun Yatsen
University (Social Science Edition), 4, 20–27.
Huang, G.-W. (2000). Systemic-functional linguistics: Forty years on. Foreign Language
Teaching and Research, 1, 15–21.
Huang, G.-W. (2003). Enhanced theme in English: Its structures and functions. Taiyuan: Shangxi
Education Publishing House.
Jespersen, O. (1949). A modern English grammar on historical principles. London: Allen &
Unwin.
Kruisinga, E. (1932). A handbook of present-day English part II. English accidence and syntax.
Groningen: P. Noordhoff.
Lock, G. (1996). Functional English grammar: An introduction to second language teachers.
Cambridge: CUP.
Matthiessen, C. M. I. M. (1995). Lexico-grammatical cartography: English systems. Tokyo:
International Language Sciences Publishers.
Morley, G. D. (2000). Syntax in functional grammar: An introduction to lexicogrammar in sys-
temic linguistics. London: Continuum.
Morley, G. D. (2004). Explorations in functional syntax: A new framework for lexicogrammatical
analysis. London: Equinox.
Morris, R. (1886[2010]). Historical outlines of English accidence: Comprising chapters on the
history and development of the language and on word formation. Charleston: BiblioBazaar.
Oliphant, T. L. K. (1878). The old and middle English. London: Macmillan.
Onions, C. T. (1905[2010]). An advanced English syntax. Whitefish: Nabu Press.
Pence, R. W., & Emery, D. W. (1965). A Grammar of present-day English (3rd ed.). New York:
The Macmillan Company.
Poutsma, H. (1929). A grammar of late modern English. Groningen: P. Noordhoff.
Quirk, R., & Greenbaum, S. (1973). A university grammar of English. Essex: Longman.
Quirk, R., Greenbaum, S., Leech, G., & Svartvik, J. (1985). A comprehensive grammar of the
English language. London, New York: Longman.
Ross, C. H. (1893). The absolute participle in middle and modern English. Publications of the
Modern Language Association of America, 8(3), 245–302.
Stump, G. T. (1985). The semantic variability of absolute constructions. Dordrecht: D. Reidel
Publishing Company.
Sweet, H. (1903). A new English grammar logical and historical part II: Syntax. Oxford:
Clarendon Press.
6 1 Introduction
2-2a.
Our guest offering his assistance, he was accepted among the number.
(The Vicar of Wakefield, 1766)
b.
The whole building being of wood, it seemed to carry every sound, like a
drum. (Women in Love, 1921)
The purpose of using absolute clauses is to avoid the subjects of the two clauses
referring to the same person or thing, and the sentences in 2-2 are rare and uncom-
mon (Onions 1905[2010]).
Latin ablative absolutes are usually translated as the “with + noun + partici-
ple” constructions in English. Influenced by ablative absolutes, traditional gram-
marians generally hold that absolute clauses are not introduced by a subordinate
conjunction, but they are always introduced by with. Here “introduced by with”
has two possible interpretations: One is that the logical subject-predicate construc-
tions following with are absolute clauses; the other is that the with constructions
themselves are absolute clauses.
10 2 Absolute Clauses in the Literature
Free adjunct constructions refer to the adverbial clauses with no explicit subject;
they are not connected with the main clauses with conjunctive expressions and are
usually separated by punctuation marks. The logical roles played by free adjuncts
2.2 Types of Absolute Construction 11
are diverse, such as time, cause, attendant circumstance, manner, result, condi-
tion, or concession (Frank 1972: 312–213; Zandvoort 1972: 37; Kane 1983: 756;
Stump 1985: 2). This construction “is native to English; examples can be found
in abundance from Old English through Modern English” (Stump 1985: 37).
Absolute adjuncts are structurally diverse, with the core components being non-
finite verbs, nouns, adjectives, adverbs, prepositional phrases, and so on, hence
nonfinite clauses in complex sentences. For example,
2-4a.
Walking home, he goes through one large garden gate, only to see the other
one fall down. (BNC_MISC)
b.
A carpenter, he had a long record of hard work, and his family was depend-
ent on his income. (BNC_NEWS)
c.
Unable to meet his eyes, she stared at the garden, wondering vaguely why it
looked the same when she felt so very different. (BNC_FIC)
d.
In his teens, he learned to drink and swore an allegiance to the pint.
(BNC_FIC)
In free adjuncts with non-finite verbs as the core components, the non-finite verbs
can be present participle, past participle, or infinitive. For example,
2-5a. Glancing up at the sky, she saw the storm clouds gathering on the horizon.
(BNC_FIC)
b. Published in 1941, it influenced many biographers. (BNC_MISC)
c. To tell you the truth, we are a bit at a loss. (BNC_FIC)
A free adjunct can also be embedded into another subordinate clause. See 2-6:
2-6 When the case, sitting at Chichester Rents in central London’s Chancery
Lane, opened nearly two months ago, Mr. Rook told the jury the fraud
involved 219 m. (BNC_NEWS)
The subjects of free adjunct constructions in the above example sentences are
all controlled by the subjects of the superordinate clauses, hence “related free
adjuncts” (Visser 1972: 1132). It is also possible that a free adjunct construc-
tion is not controlled by the subject of the superordinate clause; “in this case the
adjunct is termed unrelated or (more prescriptively) misrelated” (Stump 1985: 7),
or referred to as “dangling participle” (Visser 1972: 1132). For example:
2-7a.
Crossing the road, a lorry knocked him down. (Bailie and Kitchen 1979: 294)
b.
Playing backgammon and swapping jokes, the evening passed very pleasantly.
(Hodges and Whitten 1977: 254)
The free adjunct construction in 2-7a is syntactically related with a lorry, subject
of the main clause. In other words, grammatically, it is a lorry that is crossing the
road, but logically the subject of the free adjunct construction refers to him, object
of the main clause. In 2-7b, the free adjunct construction has no reference in the
main clause, that is, to say that the logical subject of the free adjunct construction
is implicit in the main clause, where there appears only the grammatical subject,
but this will not affect the effectiveness of communication.
12 2 Absolute Clauses in the Literature
Dangling construction has been used in all times throughout the history.
Grammarians use different names for this construction, such as loose partici-
ple (Jespersen 1949; Vallins 1952), fused participle or sentry participle (Fowler
1965: 215, 438), unattached participle, unrelated participle, pendant participle
or dangling participle (Quirk et al. 1985: 1121, 1123), and detached participle
or murky participle (Kane 1983: 255, 761). Grammarians have long been con-
cerned with dangling participles, such as Lowth (1762), Grant (1808), Cooper
(1831), and Curme (1912). Dangling participles are accepted as grammati-
cal without any criticism and “apparently not considered as irregular.” (Visser
1972: 1140).
It is Bain (1863) who first found fault in dangling construction. He “con-
demns the usage as an error arising from confounding the participle adjunct with
the absolute construction” (Visser 1972: 1140). Many other grammarians (e.g.,
Onions 1905[2010]; Partridge 1949; Fowler 1965) also question this construction,
for the logical subject and the grammatical subject of a dangling participle are not
co-referential. Although this may not cause misunderstanding, most grammarians
suggest avoiding this construction in language practice (e.g., Smart 1931; Hodges
and Whitten 1977; House and Harman 1950).
“Nowadays, in ‘literary’ English the idiom is avoided” (Visser 1972: 1140),
except occasionally appearing in premeditated spoken English, because although
it is ambiguous in structure, it is seldom misunderstood. Participles in some dan-
gling constructions have been being fixed with the gradual weakening of the
requirement of subjects, as in 2-8a, and some even are grammaticalized into prep-
ositions, as in 2-8b.
2-8a. Strictly speaking, I’ve no role here any more. (BNC_FIC)
b. Considering the interest rate blows, the market turned in a resilient perfor-
mance. (BNC_NEWS)
According to Quirk et al. (1985: 1122–1123) and Greenbaum (1996: 337–
338), the dangling construction is acceptable if it is a style disjunct that has the
speaker’s I as the understood subject, if the understood subject is a generic you,
we, or one or if it refers to the whole of the host clause, or in scientific usage,
if the understood subject refers to the I or we of the speakers or writers. For
example,
2-9a. Putting it mildly, you have caused us some inconvenience. (Quirk et al.
1985: 1122)
b. To check on the reliability of the first experiment, the experiment was repli-
cated with a second set of subjects. (ibid.: 1123)
c. Being Christmas, the government offices were closed. (ibid.: 1122)
d. Unknown to his closest advisers, he had secretly negotiated with an enemy
emissary. (ibid.: 1122)
2.2 Types of Absolute Construction 13
Absolute clauses are “nonfinite and verbless adverbial clauses that have an overt sub-
ject but are not introduced by a subordinator and are not the complement of a preposi-
tion” (Quirk et al. 1985: 1120); they are also known as “adverbial participle clauses
and adverbial verbless clauses” (Greenbaum 1996: 338). The people or thing conduct-
ing the action of the non-finite verb is not co-referential with the people or thing as the
subject of the main clause, so it is likely to be confused with a dangling construction.
The major difference between the two is that an absolute clause consists of a logical
subject and a logical predicate, while a dangling construction has no subject. Like free
adjuncts, absolute clauses are also various in structure, and their syntactical structures
can be distinguished from two dimensions, i.e., types of subject and types of predi-
cate. Modern grammarians generally accept that the subject of an absolute clause may
be a zero case noun or a nominative pronoun. They categorize the syntactic types of
absolute clauses mainly from the core component of the logical predicate, which can
be a non-finite verb or a verbless component. Non-finite verbs include present parti-
ciples, past participles, or infinitives; and verbless components include nouns, adjec-
tives, adverbs, or prepositional phrases. These structural types are shown in Fig. 2.1.
2-10a. She lay for a long while, the tears falling. (BNC_FIC)
b. Fish done, spuds on, she got out the asparagus and found they were plas-
tic! (BNC_MAG)
c. He also gave advance information about an evening meeting…and a one-
day conference…, both events to take place in London. (BNC_ACAD)
d. Her parents, Dad a solicitor and Mum a former Welsh international, are
watching with her brother as she beats Jennifer Santrock. (BNC_NEWS)
e. He dead, and you dying, he gave you the kiss of life. (BNC_SPOK)
f. Exhausted and confused, I came, cap in hand, busking for help and half
sang, half cried. (BNC_MISC)
g. Episode over, put it out of your mind. (BNC_FIC)
Grammarians generally hold that absolute clauses are not popularly used in
Modern English, “apart from a few stereotyped phrases, absolute clauses are for-
mal and infrequent” (Stump 1985: 10; Quirk et al. 1985: 1120). For example,
2-11a. Weather permitting, the big helicopters will place them between the flows
and the town. (BNC_NEWS)
b. All things considered, she would be better married. (BNC_FIC)
Present participle Past participle Infinitive Noun Adjective Adverb Preposition phrase
c. There being only one way to get out of here, she went to find Felipe.
(BNC_FIC)
d. It being Ten o’clock, the debate stood adjourned. (BNC_MISC)
The absolute constructions, i.e., free adjuncts, nominative absolutes, and aug-
mented absolutes, are all reduced finite clauses (Curme 1931: 156; Quirk and
Greenbaum 1973). The reason why they are called absolute construction is that
2.2 Types of Absolute Construction 15
there are two important similarities among them: All of them have their own tone
and are separated by a pause from the main clause, or by a comma in writing;
all of them are non-finite or verbless adverbial clauses, hence no tense or mood
marks.
However, there are also significant differences among them. Free adjuncts and
absolute clauses are not introduced by an explicit subordinator and are not depend-
ent on the main clause. This is the indicator of difference between these two and
augmented absolutes. Quirk et al. (1985: 1123) refer to these two types of absolute
constructions as supplementive clauses.
Moreover, there are two differences between free adjuncts and absolute clauses.
First, the subject of a free adjunct is implicit, and the implicit subject can be co-
referential with the subject of the main clause, hence related free adjunct or unre-
lated free adjunct. However, the subject of an absolute clause is explicit, and this
explicit subject is not co-referential with the subject of the main clause. Second,
free adjuncts can be introduced by a subordinator which can be a conjunction or
a preposition, while an absolute clause cannot be introduced by a conjunction, but
can by a preposition, such as with, forming an augmented absolute. From this per-
spective, a free adjunct is called absolute construction because there is not a sub-
ordinator connecting it with the main clause. An absolute clause is called absolute
construction, for it has not a subordinator to connect with the main clause, nor
a co-referential subject with the main clause. Free adjuncts are non-finite clauses
without explicit subordinators, and augmented absolutes are logical subject-predi-
cate constructions with explicit subordinators.
(iii) contrast: Like all craftsmen of the kind, he is at the mercy of his material.
(iv) alternative circumstances: Genuine, or a joke of the enemy, it spoke waken-
ing facts to him.
Although Kruisinga (1932: 280) has not listed the logical roles of absolute
clauses, he considers absolute clauses themselves as free adjuncts. Absolute
clauses differ from free adjuncts only in that they have their own subjects, so the
logical roles assumed by free adjuncts are the same as those assumed by absolute
clauses in his understanding.
Jespersen (1949: 61–64) thinks that it is not always easy or even impossible to
draw a clear line between several applications, but he suggests four logical roles of
absolute clauses:
(i) cause or reason: The wise men of antiquity…were afraid that—men being
what they are—their discoveries might be put to bad or futile uses.
(ii) condition: Conciliation failing, force remains; but force failing, no further
hope of reconciliation is left.
(iii) time: And the meal being over, he took Mr. Kaye into the other room.
(iv) descriptive circumstances: He remained in town, his idea being that he
wanted everything settled before his departure.
Visser (1972: 1054–1056; 1132–1139; 1149–1158; 1252–1255; 1266–1271)
discusses the range of logical roles played by free adjuncts and by absolute clauses
separately. He distinguishes four logical roles of absolute clauses: attendant cir-
cumstances; reason, ground, cause, or motive; time; and condition.
According to Quirk and Greenbaum (1972: 762), except for the attendant circum-
stances asserted by Kruisinga (1932), free adjuncts and absolute clauses have a more
specific sense in context. They suggest three logical roles played by absolute clauses:
(i) cause: All our savings gone, we started looking for jobs.
(ii) time: Cleared, this site will be very valuable.
(iii) circumstance: A case in both hands, Mabel stalked out of the house.
The above classifications of logical roles played by absolute clauses are repre-
sented in Table 2.2.
Adverbial clauses in traditional grammar can play the logical roles of time,
cause, place, condition, concession, manner, and comparison. However, there
are no adverbial clauses of attendant circumstance. Attendant circumstances are
2-18a, and the relationship of time is most possible with absolute clauses having
past participles, adjectives, or adverbial phrases in their predicates, as in 2-18b–c.
Absolute clauses of time and condition usually precede the superordinate clauses,
as in 2-18b–d, and those expressing attendant circumstances are “generally
added after the main part of the sentence” (1949: 63), as in 2-18e–g. Like Curme
(1931), Jespersen (1949) also indicates that predicates of absolute clauses express-
ing attendant circumstances are usually adverbial or prepositional phrases, some
of which are, in Jespersen’s terms, the “condensed constructions”, such as hat in
hand, head first, and face down. See the following examples:
2-18a. The terrain being flat, the wind tore across scrub and heathland unim-
peded. (BNC_FIC) (cause)
b. Dishes done, I return home to find my bucket full of ‘slime’ which I pour
into the washing machine. (BNC_NA) (time)
c. Introductions over, Nicholson motioned for his guests to sit down.
(BNC_FIC) (time)
d. Weather permitting, the big helicopters will place them between the flows
and the town. (BNC_NEWS) (condition)
e. She rushed from the shop, hat in hand. (BNC_ACAD) (circumstance)
f. Then she dragged the body to the workbench and pushed it into the cup-
board, head first. (BNC_FIC) (circumstance)
g. Bissell’s body had been found lying on a pile of rubbish, face down.
(BNC_FIC) (circumstance)
“Both semantic and pragmatic factors may determine the logical role which
adjuncts and absolutes are felt to play” (Stump 1985: 22); Quirk and Greenbaum
(1972: 762) have discussed the pragmatic factors determining the logical roles of
absolute clauses. For example, the absolute clause in 2-19a may express the rela-
tionship of cause or time and that in 2-19b, condition or cause. “For the reader or
hearer, the actual nature of the accompanying circumstance has to be inferred from
the context” (Quirk et al. 1985: 1124). For example:
2-19a. Vanity overcoming discretion, Sherman phoned the Newark Evening News
to boast of his own treasure trove. (COHA_MAG)
b. Such being the case, a few remarks will be made on each kind separately.
(COHA_NF)
Quirk and Greenbaum (1972) also offer an explanation of semantic correlate of
absolute clauses. For example, “in—ing clauses, dynamic verbs typically suggest a
temporal link, and stative verbs a causal link” (Quirk and Greenbaum 1972: 762).
This can be taken as the third factor, the semantic factor that determines the logi-
cal relationship between the absolute clause and the main clause. For example,
2-20a. Money being scarce, Belen’s people don’t buy but barter. (COCA_FIC)
(cause)
b. Sanity returning, he ran after her. (BNC_FIC) (time)
2.3 Logical Roles of Absolute Clauses 19
Syntactic, pragmatic (reasoning), and semantic factors together determine the logi-
cal roles of absolute clauses. These factors help readers make a determination on
the most appropriate role from several possible logical roles.
Logical roles distinguished by grammarians can be classified into two catego-
ries: adverbial clauses and attendant circumstances. The difference between the
two is that the former can be expanded into clausal adjuncts introduced by sub-
ordinating conjunctions and the latter, prepositional phrases introduced by with or
coordinating clauses linked by conjunction and. For example,
2-21a. Dexter turned to Emma, eyes blazing an apology. (CLOB_P)
(circumstance)
b. There being no bridge, the master had to stop at the shore. (BROWN_E)
(cause)
c. This done, she contemplated with dismay the solitary hours that lay before
her. (LOB_P) (time)
d. The unexpected weather aside, it had been a good day. (CLOB_N)
(concession)
e. All things considered, the highway commissioners would seem to be
elected. (BROWN_C) (condition)
The absolute clause in 2-21a is an attendant circumstance, and those in 2-21b–e
play the relations of cause, time, concession, and condition, respectively. Due to
the absence of explicit conjunctive expressions, the logical roles expressed by
absolute clauses are always fuzzy. Grammarians such as Curme (1931), Jespersen
(1949), and Quirk and Greenbaum (1972, Quirk et al. 1985) have discussed on
the factors determining the logical relations assumed by absolute clauses, but clear
distinctions can hardly be attained in many cases. For example,
2-22a. Her fears somewhat lulled, she began to read. (LOB_N)
b. It was a very English sort of day, the air still, the sky a uniform white.
(CLOB_L)
c. The family circle was a tight one, the discipline strict. (FLOB_G)
d. That’s twice he did it, twenty years apart, two pregnancies ending in noth-
ing, nothing. (FLOB_K)
The absolute clause in 2-22a can be interpreted as expressing the relation of
time or cause. No matter which of the two relationships it expresses, it belongs
to the category of adverbial clauses. The distinctions between different types of
logical roles are also fuzzy. For example, the absolute clause in 2-22b can be con-
sidered as an adverbial clause expressing the relationship of cause or as an atten-
dant circumstance. It is grammatically acceptable that the preposition with can be
added to such absolute clauses as time, cause, and condition to form augmented
adjuncts. Some absolute clauses may have more interpretations, as in 2-22c, the
absolute clause can be taken as an adverbial clause of cause or an attendant cir-
cumstance or neither of the two. Rather, it is the explanation to the main clause.
Although traditional grammar does not distinguish the explanation type of abso-
lute clauses, the absolute clause in 2-22d is hard to be included in the categories
20 2 Absolute Clauses in the Literature
main clauses. There is a clear distinction between 2-25 and 2-26 although they
appear the same in structure.
2-26 The Benbergs stood by, he clasping his hands and watching her closely,
she wiping a plate round and round with a sodden cloth. (Visser 1972)
According to Visser (1972), the absolute clauses in 2-26 are attendant circum-
stances. In fact, the relationships expressed by these two absolute clauses are also
to some extent fuzzy. From the perspective of predicate, the verbs in the two abso-
lute clauses are both accompanying the verb stood in the main clause. However,
from the perspective of subject, the two absolute clauses can both be seen as the
explanation of the nominal group the Benbergs, subject of the main clause. Even
so, if the main verb stood is interpreted as a state of affairs, as The Benbergs were
there, the two absolute clauses are more inclined to function as explanation.
There are two different interpretations for the case of absolute clauses to change
from dative to nominative. One believes that absolute clauses originated from
dative absolutes in the Old English, which were borrowed from Latin ablative
absolutes. According to this interpretation, the dative case in Anglo-Saxon is the
origin of the absolute clauses in English, and the absolute case changed from
dative to nominative owing to the loss of case inflections. “The inflections having
decayed, the dative was mistaken for the nominative” (Kellner 1892: 125). On the
change of absolute case from dative to nominative, Bright (1890: 159–162) wrote
the following:
Let us look at the history of the absolute construction in English. We begin with the
dative absolute in Anglo-Saxon (in origin a translation of the Latin ablative absolute); as
inflections break down we come upon the transition or ‘crude’ type, in which the pro-
noun remains dative in form while the participle has lost all signs of inflection. But all
nouns, as well as the participle, came to lose the inflectional sings of the dative case; we
then obtained the ‘crude’ type, in which both noun and participle, though absolute, were
without any trace of inflection. The final act in this history was the admission of the nom-
inative forms of the personal pronouns into this crude absolute construction – a dative
absolute in disguise.
Another interpretation believes that the use of a noun in the zero form or a
pronoun in the subject form is “a continuation of the Old English usage with the
noun before the participle in the zero case, with later analogous introduction of
the subject form of the pronouns. This latter phenomenon took probable place
as early as the middle of the fourteenth century” (Visser 1972: 1149). According
to this interpretation, the subject of absolute clauses should always be nomina-
tive because they evolved from the Old English. An absolute clause “is errone-
ous in making it the objective” (Murray 1808[2011]: 201), and Lowth (1762: 116)
warned against dative absolute in disguise “forcing of the English under the rules
22 2 Absolute Clauses in the Literature
c. With a new formality and silence she led the way into the hall, he following.
(1894)
d. She had turned back to the drawing-room, forgetting the other guests, he
walking beside her. (1894)
e. The Benbergs stood by, he clasping his hands and watching her closely, she
wiping a plate round and round with a sodden cloth. (1894)
f. She being down, I have the placing of the British crown. (1894)
It is noteworthy that accusative pronouns as subject of absolute clauses com-
pletely disappeared after the fifteenth century and reappeared in informal English
in the nineteenth century (Visser 1972: 1147). There are two possible interpreta-
tions: One is that absolute clauses with accusative pronouns as subject existed in
spoken English in this period of time; the other is that the use of accusative pro-
nouns in Modern English is the same as that of such expressions as it is me, that’s
him, etc. in origination. According to Burn (1766[2010]: 61), despite the fact that
2-29 is not correct in grammar, it is still necessary to speak like that.
2-29 Him watching, all the rest went to repose themselves. (1766)
Bain (1904: 273) found that the accusative form corresponding to the dative
form in the oldest English is not unusual until recently. For example,
2-30a. But you see, him being here, in the room—I had to be careful. (1926)
b. It made me so tired, it did. Him worshipping the ground she trod and her
not caring a snap of the fingers for him. (1932)
c. You’ve had a disappointment, I Know, her being away. (1933)
Grammarians have paid full attention to the case of absolute clauses, but they
have not reached any agreement so far. In actual language use, absolute clauses
can be nominative, as in 2-31, or accusative, as in 2-32.
2-31a. Of course he was thirteen years older than her, she being but twenty.
(BNC_FIC)
b. Some twenty thousand people attended to hear him speak, I being one of
them. (BNC_MISC)
c. He whispering endearments in his lover’s ear, the joy of lying in each oth-
er’s arms… (BNC_FIC)
2-32a. She was surprised he lived so poorly, him being a successful man.
(BNC_FIC)
b. Boys pouring into the room below, laughing, chattering, me seeing them
through the crack. (BNC_FIC)
c. You asked her if she would like to have the bairn, her being a minister’s
wife. (BNC_FIC)
24 2 Absolute Clauses in the Literature
In the first half of the Middle Ages, absolute clauses were practically non-existent.
As Ross (1893: 296) states:
Its prevalence in Chaucer is due largely to Italian influence, in part also to French influ-
ence, and the occurrence of the participle in the works of Chaucer’s contemporaries and of
the fifteenth century writers is to be traced to the same French influence. But the construc-
tion was avoided as much as possible, and in its stead the various shifts that were resorted
to in Anglo-Saxon were used. The absolute participle here cannot be spoken of as “a norm
of style”… During the fifteenth century, however, just before the awakening caused by the
Revival of Learning, the absolute participle became, as we have seen, somewhat prevalent
and was more felt in the style. (Ross 1893: 296)
In the Modern English period, absolute clauses have assimilated and developed
into a style and were used by all writers. In fact, early in the second half of the
seventeenth century, absolute clauses became completely natural. They became
thoroughly fixed as a style in the eighteenth century. At this time, the appearance
of the novel as a style makes the use of absolute clauses very popular.
Absolute clauses are used differently in different styles of works. “In the six-
teenth and seventeenth centuries it belonged largely to didactic and philosophi-
cal prose, but now its province is distinctively narration and description…Next
to this stand biography, history, and the essay” (Ross 1893: 297–298). In the
Anglo-Saxon and medieval English, absolute clauses were nearly used only in
prose, rarely in poetry. Chaucer is an exception. This is because Chaucer imi-
tated Boccaccio in whose poems there are a large number of absolute clauses.
In Shakespeare’s poems absolute clauses are rarely seen, but very popular in the
poems of Dryden and the poets of the eighteenth and nineteenth centuries.
Absolute clauses were severely inveighed by grammarians of the nineteenth
century. “Even such forms as ‘Herod being dead, the angel warned Joseph’ seem
rare in the best recent English” (Mcelroy 1885: 105). “The participial construction
is a convenient means of condensation… Being, however, a subordinated construc-
tion, it needs careful adjustment to the principal assertion on which it depends”
2.5 Stylistic Effects of Absolute Clauses 25
(Genung 1885[2010]: 117). “This construction belongs to literary style rather than
to colloquial speech” (Jespersen 1933: 313).
In general, absolute clauses are a good choice which can give sentences life and
movement and make sentences free and diverse. “Because nonfinite clauses lack
tense markers and modal auxiliaries and frequently lack a subject and a subordi-
nating conjunction, they are valuable as a means of syntactic compression. Certain
kinds of nonfinite clause are particularly favoured in written prose” (Quirk et al.
1985: 995). In Modern English, absolute clauses have become an important part of
English syntax, forming a particular style. They are used in the works of all writ-
ers, both in literary works and in oral conversation, even though mainly confined
to some fixed expressions, such as all things being equal and all things consid-
ered. However, Modern grammarians (e.g., Quirk et al. 1985) generally hold that
absolute clauses are normal and infrequent, mainly used in formal works and sel-
dom in spoken or informal texts. Jespersen (1949: 62) illustrates the situations of
application of absolute clauses through 2-33 as follows:
2-33a. He stood, hat in hand.
b. He stood, his hat in his hand.
c. He stood, with his hat in his hand.
Obviously, a search of similar phenomenon in corpora reveals that 2-33b is
rarely seen, 2-33c is commonly used in spoken language, while 2-33a is relatively
common in literary works.
2.6 Questions to Be Answered
different from the main clauses in position. Quirk et al. (1985: 1123) distinguish
the positions of attributive clauses and adverbial clauses with their main clauses.
2-35a. Jason, told of his son’s accident, immediately phoned the hospital. (Quirk
et al. 1985: 1123)
b. Jason, who was told of his son’s accident, immediately phoned the hospi-
tal. (ibid)
The position of adverbial clauses is arbitrary. They can be positioned ini-
tially, medially, and finally, while the most typical position of attributive clauses
is immediately after their antecedent. If subjectless nonfinite clauses occur in that
position, they may be indistinguishable from the participle clauses functioning as
post-modifiers or noun phrases in apposition. See 2-36 below:
2-36a. This substance, discovered almost by accident, has revolutionized
medicine.
b. This substance, which was discovered almost by accident, has revolution-
ized medicine.
c. Discovered almost by accident, this substance has revolutionized
medicine.
The non-finite clause in 2-36a can be interpreted as a post-modifier, as in
2-36b, or as a subjectless non-finite adverbial clause, as in 2-36c. The absolute
clauses in 2-34 are positioned after the antecedent and they cannot be positioned
before the main clause, hence not adverbial clauses but appositive clauses. The
same is true for the absolute clauses in 2-26.
According to the analysis above, it is problematic to define absolute clauses as
non-finite and verbless adverbial clauses with an explicit subject but without an
introducing subordinator, because being adverbial is not the necessary require-
ment for constituting absolute clauses. Absolute clauses are not always non-finite
or verbless adverbial clauses with an explicit subject; they may also be non-finite
or verbless appositive clauses with an explicit subject. Accordingly, at least three
questions in research available need to be answered.
Question One Since being adverbial is not the necessary requirement for form-
ing absolute clauses, then what are the identification criteria for
absolute clauses?
Question Two Since appositive clauses with an explicit subject can form abso-
lute clauses, then can other function types of nonfinite or verb-
less clauses with an explicit subject form absolute clauses?
Question Three What are the characteristics of historical, stylistic and case dis-
tribution of absolute clauses?
Many grammarians believe that absolute clauses are formal in style and are
decreasing in actual use, but the stylistic distributions of absolute clauses or the
function types of absolute clauses have not been explored in detail. Scholars dis-
tinguish different logical roles played by absolute clauses, but many problems still
remain. Deficiencies of current studies are obvious. In SFL, absolute clauses are
2.6 Questions to Be Answered 27
almost totally ignored too. Only when discussing non-finite clauses are absolute
clauses mentioned, i.e., “there may be an explicit Subject in the dependent clause”
(Halliday 1994: 229; Halliday and Matthiessen 2004: 404). See 2-37 below:
2-37a. John went off by himself, the rest of us staying behind.
b. It’s a much bigger house, for the children to have their own rooms.
We may simply say that, the non-finite clause in 2-37a, the rest of us staying
behind, has an explicit subject which does not refer to the subject of the primary
clause, hence is an absolute clause; and the nonfinite clause in 2-37b, for the chil-
dren to have their own rooms, has also an explicit subject, but it has an explicit
conjunctive preposition, hence is not an absolute clause. This is far from SFL anal-
ysis. More importantly, it is not readily applicable to many other instances.
In the next chapter, we will first offer a sketch of systemic functional theory,
and then discuss the SFL approach to absolute clauses.
References
Adams, E. (1858[2010]). The elements of the English language. Whitefish: Kessinger Publishing.
Al-Hamash, K. I., & Abdulla, J. J. (1968). A course in English linguistics. Baghdad: Al-Sha’b
Press.
Al-Hamash, K. I., & Abdulla, J. J. (1979). A course in modern English grammar. Baghdad:
IDELTI.
Annema, H. (1924). Die sogenannten absoluten participalkonstructionen im neuhochdeutschen.
Groningen: de Waal.
Bailie, J., & Kitchin, M. (1979). The Hamlyn guide to English usage. London: Hamlyn.
Bain, A. (1863). An English grammar. London: Longman.
Bain, A. (1904). A higher English grammar. London: Longman.
Berent, G. B. (1973). Absolute constructions as “subordinate clauses”. In C. Corum, C. Smith-
Stark, & A. Weiser (Eds.), You take the high node and I’ll take the low node: Papers from the
comparative syntax festival (pp. 147–154). Chicago: Chicago Linguistics Society.
Berent, G. B. (1975). English absolutes in functional perspective. In R. E. Grossman, L. J. San,
& T. J. Vance (Eds.), Papers from the parasession on functionalism (pp. 10–23). Chicago:
Chicago Linguistics Society.
Beukema, F. H. (1980). Prepositions and tenseless sentence adverbials in English. In S. Daalder
& M. Gerritsen (Eds.), Linguistics in the Netherlands 1980 (pp. 69–79). Amsterdam, Oxford
& New York: North-Holland.
Beukema, F. H. (1982). On the internal structure of free adjuncts. In S. Daalder & M. Gerritsen
(Eds.), Linguistics in the Netherlands 1982 (pp. 71–82). Amsterdam: North-Holland.
Biber, D., Johansson, S., Leech, G., Conrad, S., & Finegan, E. (1999). Longman grammar of spo-
ken and written English. London: Longman.
Bright, J. W. (1890). The objective absolute in English. Modern Language Notes, 3, 159–162.
Brown, G. (1861). The grammar of Englsih grammars (6th ed). New York.
Burn, J. (1766[2010]). A practical grammar of the English language. Charleston: BiblioBazaar.
Callaway, M. (1889). The absolute participle in Anglo-Saxon. Baltimore: Press of I. Friedenwald.
Cooper, J. G. (1831). A plain and practical English grammar. Philadelphia: Judah Dobson.
Crystal, D. (2008). A dictionary of linguistics and phonetics (6th ed.). Oxford: Blackwell
Publishing.
Curme, G. O. (1912). A history of the English relative construction. London: Longman.
Curme, G. O. (1931). A grammar of the English language: Syntax. Boston: D. C. Heath.
28 2 Absolute Clauses in the Literature
3.1 Introduction
Like traditional syntax, functional syntax consists of two parts: form and func-
tion. Traditional syntactic analysis is a maximal bracketing approach, i.e., imme-
diate constituent analysis (IC analysis), concerning about what meanings a form
can express; functional syntactic analysis is a minimal bracketing approach, i.e.,
ranked constituent analysis, concerning about what forms can realize a certain
meaning. Traditional grammar constitutes a set of syntactic structure, while func-
tional grammar, three sets of syntactic structure. For example, the syntactic struc-
tures of example 3-1 are shown in Fig. 3.1:
3-1 Shortages made life difficult, to the point of exasperation. (BNC_NA)
Traditional grammar includes absolute clauses into the category of non-finite
and verbless adverbial clauses with an explicit subject. They have their own sub-
ject–predicate structures. According to SFL, absolute clauses are at the clause rank
in rank scale. Being clauses, they can be analyzed in transitivity structure, mood
structure, thematic structure, and information structure.
3.2.1 Transitivity Structure
Transitivity Goal
Actor Process: Material
Structure Carrier Attribute
processes, and circumstances. The process is the core element, realized by verbal
groups, and participants are close to this core element, realized by nominal groups.
Therefore,
We can say that the configuration of process + participants constitutes the experiential
centre of the clause. Circumstantial elements augment this centre in some way — tempo-
rally, spatially, causally and so on; but their status in the configuration is more peripheral
and unlike participants they are not directly involved in the process. Circumstances are
realized by adverbial phrases or prepositional phrases. (Halliday and Matthiessen 2004:
176)
SFL distinguishes six types of process. Non-finite verbs realizing the six types
of process can all constitute absolute clauses. In the absolute clauses constituted
by verbless clauses, the relational verbs are omitted, with minimal groups, adver-
bial groups, or prepositional phrases realizing attribute. For example,
3-2a. Hands shaking, I walked deeper into the house, Benjamin behind me.
(BNC_FIC) (Material)
b. The Lord being good, Joseph hopes for a new son for the House of David.
(COCA_FIC) (Relational)
c. It’s always been security problems, people talking about war. (COCA_
NEWS) (Verbal)
d. Tiara sat, people watching, as she sipped her punch. (COCA_FIC)
(Behavioural)
e. Jo Ellen could hardly stand it, people knowing her name, asking her for rec-
ipes, hugging her. (COCA_MAG) (Mental)
f. I went upstairs to my room, there being no other place. (COCA_FIC)
(Existential) (Fig. 3.2)
3.2.2 Mood Structure
Interpersonal function is the function to enact the social process or human rela-
tionship using language. “The principal grammatical system is that of MOOD.”
(Halliday and Matthiessen 2004: 106) The mood structure consists of the mood
and the Residue. Mood is the constituent formed by Subject and Finite, and the
main verb falls in the Residue. The mood structure of absolute clauses comprises
the Subject and the Residue, there being no Finite. There is no mood in a clause
without Subject and Finite, the Finite playing the decisive part. This is to say that
absolute clauses are clauses without mood. For example,
3-3 Time permitting, we shall discuss the matter. (COCA_FIC)
The absolute clause Time permitting in 3-3 contains a Subject and a main verb,
but not a Finite, hence no mood. In interrogative clauses, the Finite is the element
that can be moved before the Subject. In Time permitting, it is the Finite does or is
that is omitted. The finite clause is Time is permitting or Time permits (Fig. 3.3).
34 3 Approaching Absolute Clauses …
hands Shaking,
I walked deeper into the house, Benjamin behind me.
Medium Process: Material
people watching,
Residue
Mayhew having concluded his narration, Ahab put such questions to him.
Theme Rheme
Given New
clause complex is the initiating clause, which adopts the tone 3, indicating that
there is still other information following, and the second clause is the continuing
clause, which adopts the tone 1, indicating that the sentence ends here. However,
two independent clauses are cohesive, both adopting tone 1. In hypotaxis, the two
clauses are not equal in status. Only one of the two clauses, the dominant clause,
can be independent from the other, and the dependent clause functions as the
modification of the primary clause. The dominant clause construes a proposition
and can have a tag question, but the dependent clause cannot. A hypotactic clause
complex has no corresponding cohesive equivalents. However, two independent
clauses can realize the logico-semantic relationships of time, cause, condition,
and concession with no conjunctions realizing the relator, but sometimes there are
conjunctive adverbs to connect the two clauses. Because of the unequal status, the
position of the dependent clause depends on the requirement of the textual con-
struction; it can be positioned before or after the dominant clause. The dependent
clause positioned before the dominant clause is represented as the Theme of the
clause complex, which is made prominent through topicalization, and adopts the
tone 4 or tone 1, while the dependent clause positioned after the dominant clause
has no topic status and adopts the tone 1.
In expansion, the secondary clause elaborates, extends, or enhances the pri-
mary clause, and in projection, the primary clause projects a locution or an idea.
“Expansion relates phenomena as being of the same order of experience, while
projection relates phenomena to phenomena of a higher order of experience
(semantic phenomena—what people say and think).” (Halliday and Matthiessen
2004: 377) Interdependency and logico-semantic relations intersect to form a rela-
tion network of clause complexes, as shown in Fig. 3.5.
According to the interdependency and logico-semantic relation, Halliday
(1994: 220) distinguishes ten basic relationships of absolute clauses. See Table 3.1
In hypotaxis, “the dominant element is free, but the dependent element is not”
(Halliday 1994: 221). The semantic types of the hypotactic clause in the clause
parataxis idea
clause projection
LOGICO- locution
elaborating
SEMANTIC TYPE
extending
expansion
stop enhancing
RECURSION
go on
3.3 Functional Structure of Clause Complex 37
Although non-finite clauses are at the same rank as clauses, they have some certain
specific features different from finite clauses in syntactic function. For example,
the -ing forms of verb include present participle and gerund, and present partici-
ples share more features of clauses, while gerunds share more features of nominal
groups. In this section, we will discuss the rank status of absolute clauses accord-
ing to the theory of cline of SFL.
3.4 Rank Status of Absolute Clauses 39
3.4.1 Cline
Cline refers to those language units which cannot be included in any language
category (Halliday and Matthiessen 1999). “To categorize linguistically is to put
together the linguistic symbols and the human experiences these symbols rep-
resent so as to set up categories.” (Yang 2007: 50) The concept of “category”
originates from the Aristotelian Classical Philosophy, according to which cat-
egory is defined in terms of sufficient and necessary features, and all members
in a category are equal in status. Influenced by Wittgenstein’s thinking of family
resemblance, cognitive linguistics has raised the prototype theory of semantic cat-
egorization, pointing out that there are degrees of membership based on degrees of
similarity. Unlike the classical theory and the cognitive theory, the theory of SFL
is not oriented toward philosophy or logic, but toward language and language use,
in other words, concerning “with how meaning is construed in naturally occurring
text” (Halliday and Matthiessen 1999: 72).
According to SFL, there are no neat boundaries between categories, and
the members of the two categories form a cline. For example, “the distinction
between closed system patterns and open set patterns in language is in fact a
cline” (Halliday 1961[2002]), and that between participants and circumstances is
also a cline (Halliday and Matthiessen 1999) and the nominal groups in preposi-
tional phrases are indirect participants (Halliday 1994: 150). The notion cline is a
major indicator to distinguish SFL from formal grammars, and cline can be seen
in many concepts of SFL. For example, “material, mental, and relational processes
are the main types of process in the English transitivity system” (Halliday and
Matthiessen 2004: 171), and on the borderlines between them are the behavioral,
verbal, and existential processes. These six processes form a cline.
The idea of cline is helpful to explain many linguistic phenomena that formal
grammars have difficulty explaining. For example, influenced by Aristotle’s cat-
egory theory, people distinguished various language units and developed many
rules to categorize language. Take adverbial participles as an example. According to
traditional grammar, the subject of adverbial participles should be the subject of the
main clause, as in 3-10a; if not, it should have its own explicit subject, hence abso-
lute clauses, as in 3-10b. However, there are also some adverbial participles with the
subject not being that of the main clauses, hence dangling participles, as in 3-10c.
3-10a. Walking in the streetbeside Marla, Toni felt something again. (COCA_FIC)
b. Weather permitting, the bar is open from 6 p.m. to midnight. (COCA_NEWS)
c. Judging from Roy’s earnest manner, this was serious stuff .(COCA_FIC)
Traditional grammar refers to the participial phrases in 3-10a and 3-10c as free
adjuncts, the former being related free adjunct and the latter, unrelated free adjunct
(dangling participle). Aristotle’s category theory cannot distinguish judging from
walking, and the prototype theory of cognitive science considers judging as the
non-core member of a category and therefore has to give a semantic explanation
to this phenomenon. According to formal grammar, judging has no logical subject
because it has become a fixed phrase.
40 3 Approaching Absolute Clauses …
The notion of cline of the SFL category theory can offer a reasonable explana-
tion to judging in 3-10c. Participial phrases constitute non-finite clauses, realiz-
ing circumstances. However, the core members of the category of circumstances
are prepositional phrases. It is a cline between non-finite clauses and preposi-
tional phrases. Phrases can fall into the group rank or the word rank (Yang 2001).
Generally, verbs are in requirement of a subject more than prepositions and tend
to be grammaticalized into prepositions. Grammaticalization “is confined to the
development from lexical to grammatical forms” (Heine and Kuteva 2004: 4). For
example, considering, regarding, and concerning in English have all grammatical-
ized from verbs to prepositions. The English preposition during originates from
the old French verb durer (meaning to continue), which was loaned into English
in the Middle Ages, the present participle form being duren. Now, the word has no
traces of a verb and has completely grammaticalized into a preposition. Therefore,
it is a cline from participles to prepositions. For example,
3-11a. Looking at the picture, I could feel his love from far away. (COCA_FIC)
b. Considering the times, it seemed destined for success. (COCA_NEWS)
c. During the war, everything here went to pieces. (COCA_FIC)
The participle looking in 3-11a is the strongest in motion, and looking at the
picture is a non-finite clause. It is in strong requirement of a subject. In 3-11c, dur-
ing is the weakest in motion, and During the war is a prepositional phrase. It is not
in strong requirement of a subject. In 3-11b, considering lies in between looking
and during in motion. It is an internal mental activity and has no obvious external
relationship with the subject of the main clause.
There are also many clines in the logico-semantic relations of non-finite
clauses. For example,
3-12a. I worked for a local firm at that time, selling office equipment. (Halliday
and Matthiessen 2004: 404)
b. She lay awake for some time, puzzling over Fen’s behaviour. (BNC_FIC)
c. I’ve seen eight people leaving the Stenness Hotel. (BNC_SPOK)
3-12a is a clause complex, realizing elaboration. 3-12c is a simple clause with an
embedded non-finite clause, realizing modification. 3-12b lies in between elabora-
tion and embedding. If it is considered as elaboration, the whole structure is a clause
complex; if it is considered as embedding, the whole structure is a simple clause.
3.4.2 Rank Status
Non-finite clauses are at the clause rank. However, they have some specific fea-
tures different from finite clauses, e.g., they do not realize parataxis in clause com-
plexes. Despite the fact that both traditional grammar and SFL ascribe absolute
clauses into the category of non-finite clauses, they are not mere non-finite clauses
with a subject. They are different from non-finite clauses in that they do not need
3.4 Rank Status of Absolute Clauses 41
conjunctive expressions to realize the relator, and they are not co-referential with
the subject of the main clauses. In this section, we will discuss the rank status of
absolute clauses in terms of the notion of cline.
Halliday (1985: 193, 1994: 216) classifies lexico-grammatical units into
five basic constituents in rank, including clause complex, clause, phrase/group,
word, and morpheme. However, he later excludes clause complex from the rank
scale, leaving only four constituents: clause, phrase/group, word, and morpheme
(Matthiessen and Halliday 1997[2009]: 71–72). This is because the clause com-
plex is not a lexico-grammatical unit, and the four grammatical units can all form
complexes (Yang 2003: 44; Halliday and Matthiessen 2004: 9). There are not dis-
tinctive boundaries between the four ranks; rather, they form a cline. For example,
at the group/phrase rank, groups are bloated words and are closer to words in rank,
and phrases are compacted clauses and are closer to clauses in rank. Although
non-finite clauses are at the clause rank, their core element, non-finite verbs, deter-
mines that they are in between finite clause and group in the cline of rank scale.
Non-finite verbs distinguished by traditional grammar include present parti-
ciple, past participle, infinitive, and gerund. SFL does not make such a distinc-
tion, but points out in discussing hypotaxis that “the dependent clause often has an
explicit Subject of its own; when this can show a contrast in case, it appears either
in oblique (e.g. him) or in possessive (e.g. his) form” (Halliday and Matthiessen
2004: 421). For example,
3-13 With him/his taking time off, everyone has to work harder. (ibid)
According to the traditional grammar, in such a structure with him taking time
off, him is the object of the preposition with, and taking time off is the comple-
ment of him, forming a logical subject–predicate structure, taking being a present
participle. In with his taking time off, his taking time off is the object of with, form-
ing a prepositional phrase, taking being a gerund. “One problem of terminology
and analysis arising in connection with both free adjuncts and absolutes is that
of distinguishing present participles from gerunds.” (Stump 1985: 11) “Halliday
appears to treat the ‘conjunctive preposition’ as a conjunction for the purposes of
the clause analysis.” (Butler 2003: 269, 271) In with him taking time off, him tak-
ing time off is a non-finite clause, and with is a conjunctive preposition, realizing
relator. In with his taking time off, the possessive pronoun his “reflects the earlier
status of these non-finite clauses as rank-shifted” (Halliday and Matthiessen 2004:
421). In his taking, taking is the nominalization of take, closer to group/phrase in
rank. Therefore, his taking is a constituent of the finite clause, and it should not be
considered as a hypotactic clause in a clause complex. However, it still construes a
figure at the semantic level and construes a sequence with the finite clause. Let us
see example 3-14:
3-14a. When he arrived, they would figure out something together. (COCA_FIC)
b. The proper season arriving, the vine is again pruned, and again eight or
ten times as many buds are retained as the plant can nourish. (COHA_NF)
c. When arriving at the station he asked for two tickets.(COHA_MAG)
42 3 Approaching Absolute Clauses …
d. Arriving at a suitable pitching site, I would erect the tent and Pete would
fetch water and have a brew ready by the time we piled in. (BNC_MISC)
e. With winter arriving, they turned up one chilly night at 17th Street and
Columbia Road NW… (COCA_NEWS)
f. On arriving at the lounge, he saw Brother Emil Sanger already seated at
the far end of the circle of chairs. (COCA_FIC)
g. Upon my arriving there the Gypsies swarmed out from their tents.
(COHA_NF)
h. On arrival at Honey Cottage, Yanto introduced the two girls. (BNC_FIC)
i. On his arrival at Leipzig in the autumn of 1865, … he was in low spirits.
(BNC_ACAD)
The hypotactic clause in 3-14a is finite, conjunction when realizing the relator.
The hypotactic clause in 3-14b is an absolute clause, without a conjunctive expres-
sion realizing the relator and the logico-semantic relation being explicit and arriv-
ing is a present participle, the logical subject being a zero case noun. Even if there
is a conjunctive expression, as in 3-14e, the non-finite clause with subject func-
tions as the complement of with, and arriving is still a present participle. In 3-14c,
the hypotactic clause is non-finite, and the use of conjunction when indicates that
arriving is a present participle. In 3-14d, there is no conjunctive expression real-
izing the relator, and it may be a conjunction, as in 3-14c, or a preposition, as in
3-14f. The preposition on in 3-14f realizes relator, and the non-finite verb arriving
functions as the complement of the preposition on and this means that arriving has
begun to be nominalized. The arriving in 3-14g is modified by the possessive pro-
noun my and is further nominalized. The arrival in 3-14h and 3-14i is the nomi-
nalization of verb, constituting a preposition phrase with on. Although arrival is a
nominal group, it still construes a figure at the semantic level.
The verbal group arrive in 3-14 is gradually nominalized from 3-14a to 3-14i,
and in the process of nominalization, the degree of grammatical metaphor is
increasing. Therefore, from the point of rank scale, the nominalization process
from 3-14a to 3-14i is a downward rankshift from clause to group/phrase. Verbs
are the core elements of clauses, and they determine the rank status of clauses;
nouns are the core elements of nominal groups, and they determine the rank status
of the nominal groups. The cline of rank from verbal group to nominal group is
shown in Fig. 3.6.
This means that absolute clauses should be positioned between finite clauses
and non-finite clauses introduced by a conjunction along the cline of rank. This
analysis is also true for verbless clauses.
gerund
The purpose of this section is to probe into the requirement for forming absolute
clauses and then to find out the criteria for identifying absolute clauses from the
SFL perspective.
3.5.1 Formation Requirement
device, and it functions to make the two clauses cohesively linked together at the
textual level.
In addition, the identification of absolute clauses in the traditional sense is
writing based. An absolute clause is always separated from the other part of the
sentence by a punctuation mark. However, the conjunctive adverb however in the
above example 2-1a will still be an absolute constituent even if it is not separated
by a comma. The reason is that whether there is a comma or not, an absolute con-
stituent is not an immediate participant of the relevant clause, and the comma is
only an explicit indicator. For example, the comma in example 2-1b cannot be
left implicit because if there is no comma, the adjective alone may be mistaken to
function as the modifier of Sara although this is ungrammatical since the adjective
alone itself cannot assume the role of modifier. On the other hand, the adverb is
flexible in a clause, and it is always separated from other constituents by a comma.
Whether it is separated or not, the syntactic function of the adverb will not change.
Therefore, it cannot be taken as the identification principle or the formation
requirement of absolute clauses to say that absolute clauses are “not grammati-
cally tied to the sentence” (Kane 1983: 754) or “not explicitly bound to the matrix
clause syntactically” (Quirk et al. 1985: 1120). Absolute clauses can be defined in
the framework of SFL as follows:
Non-finite clauses consisting of a nominal group that is not a direct participant in the primary
clauses and a nonfinite phrase not introduced by an explicit conjunctive expression.
3.5.2 Identification Criteria
3.5.2.1 Relator
between the two clauses but also make the logico-semantic relation explicit. The
other is that the logico-semantic relation between the secondary clause functioning
as participant and the primary clause is unique, and it does not need to be real-
ized by explicit conjunctions, the clause functioning as complement being a pro-
jected one. The classification of projection is not realized by conjunctions but by
projecting verbs. Clauses functioning as subject are embedded clauses, which can
be introduced by the conjunction that or not. In general, finite clauses realizing
circumstances or participants are linked by conjunctions. When the finite clauses
realizing circumstances are reduced to non-finite clauses, the conjunctions can be
left implicit on condition that the logico-semantic relation is clear. The only inter-
pretation for the explicit conjunctions is that they can make the logico-semantic
relation clear. The logico-semantic relation between a non-finite clause realizing
participants and the primary clause cannot cause misunderstanding; hence, an
explicit conjunction is no longer required. Non-finite clauses in 3-18a, 3-18b, and
3-19a can all have their finite equivalents. For example,
3-18a. He heard (that) Craig was chuckling inside his helmet.
b. They consider (that) Christmas is a pagan holiday.
3-19a. (That) Tom flying planes badly can be dangerous.
The non-finite clauses in 3-19b and 3-17 also have their finite equivalents,
introduced by relative pronoun which or who. The relative clause in the former is
nonrestrictive and that in the latter, restrictive. For example,
3-19b. People who live nearby were told to stay indoors.
According to SFL, the nonrestrictive relative clause rankshifts to function as
the modifier of the nominal group, the whole construction being a simple clause,
and that in 3-17 is a hypotactic clause, elaborating the primary clause, the whole
construction being a clause complex. Therefore, having finite equivalents can be
taken as a criterion for identifying absolute clauses. So we have the first criterion
as follows:
Criterion One: Absolute clauses have their finite equivalents introduced by
conjunctions.
3.5.2.2 Participant
of the verbal group shall discuss of the primary clause. The subject one of the
absolute clause of elaboration is substitute, realizing a part-whole relation with the
antecedent boxes. In 3-17b, the subject a girl and a man of the absolute clauses
of elaboration can be seen as the complement of the verbal group of the primary
clause semantically, but they both are not the direct participants of the verbal
group of the primary clause.
The participant oriented criterion for identifying absolute clauses can be helpful
for better explaining the absolute clauses in 3-18 and 3-19. “When a noun phrase
intervenes between the host verb and the to-infinitive, it is often unclear whether
the phrase belongs to the host clause or the complement clause. In either case, if it
is a pronoun it is in the objective case.” (Greenbaum 1996: 350) 3-18a is a clause
complex of projection, in which Craig chuckling inside his helmet is a non-finite
clause with subject, the whole construction functioning as the complement of the
main verb heard. In this construction, Craig functions both the complement of
the main verb and the subject of the non-finite verbal group. Therefore, the whole
sentence is a combination of two clauses. From this sense, the subject Craig is
not independent from the main verb. However, semantically, what He heard is
the sound of laughing of Craig, rather than Craig himself. The subject Craig of
the non-finite clause cannot be separated from the non-finite verbal group; they
together realize the phenomenon of the mental verb heard. Jespersen (1933) refers
to the structure that can form a complete piece of communication as independent
nexus, which “forms only a part of a sentence, and thus may be either a primary in
a sentence (subject or object), a secondary (an adjunct) to a primary in a sentence,
or a tertiary in a sentence” Jespersen (1933: 309). Jespersen (1933: 310) explains
the simple nexus functioning as the complement of the main verb in 3-18 more
clearly through 3-21.
3-21 I found her gone. (Jespersen, 1933: 310)
It is clear that in 3-21 what I found is her gone. I did not find her because she
was gone.
3-18b and 3-18a are the same in form, but their syntactic structures are obvi-
ously different. In Halliday’s thinking, the construction considers Christmas to be
is a verbal group realizing the relationship of cause, hence not a clause complex.
However, Yang (2003) holds that both 3-18a and 3-18b are clause complexes of
projection. Both Craig chuckling inside his helmet in 3-18a and Christmas to be a
pagan holiday in 3-18b are projected non-finite clauses, the former being a macro-
phenomenon, construing an act and the latter, a metaphenomenon, realizing a fact.
Huddleston and Pullum (2002) try their effort to prove that in 3-18b Christmas is
the object of the verb consider, and to be a pagan holiday is the complement of
Christmas. The reason is that the whole construction will no longer be an accept-
able sentence when Christmas is removed. A complement is not an essential com-
ponent of a sentence. Since to be a pagan holiday is a complement in traditional
grammar, it is not indispensible. For example, He heard Craig in 3-18a is accept-
able, but They consider Christmas in 3-18b is not. The finite equivalent of the non-
finite clause in 3-18b is They consider (that) Christmas is a pagan holiday, from
3.5 Formation Requirement and Identification Criteria of Absolute Clauses 49
which it can be seen clearly that Christmas is a part of the hypotactic clause. In
fact, Christmas in 3-18b is not indispensible of consider, but is indispensible of to
be a pagan holiday. In addition, the nominal group Christmas following consider
cannot be omitted because the subject of the non-finite clause is not co-referen-
tial with that of the primary clause and it requires an explicit subject of its own.
However, even if the two subjects are co-referential in such a construction, the
subject of the non-finite clause is still indispensible, and then, a reflexive pronoun
like himself is required to fill the subject slot. Despite the fact that reflexive pro-
nouns have always an antecedent item, they themselves always carry an additional
meaning of emphasis, which prevents them from complete co-referential with the
antecedent. Nominative or accusative pronouns can be completely co-referential
with the antecedent, so they are acceptable grammatically and cannot form abso-
lute clauses.
The projected non-finite clause can be a fact or an act. If it is a fact, even if the
subject of the non-finite clause can be the complement of the main verb of the pro-
jecting clause, the projected clause still forms an absolute clause. This is because
a verb projecting a fact and a verb having a nominal group functioning as its com-
plement are not different in meaning. For example, the main verb believes in 3-22
means “to accept as true,” which is different from that in she believes Nathan,
meaning “trust.”
3-22 She believes Nathan to be a fine young man. (COCA_FIC)
Structurally, 3-22 is closer to 3-18a than 3-18b is. The difference between
3-18a and 3-22 lies in that in the former the secondary clause is a project act,
while in the latter, a projected fact. This can be tested by changing the non-
finite clauses into relative clauses. For example, 3-18a can be changed into He
heard Craig, who is chuckling inside his helmet, without changing the meaning.
However, if 3-22 is changed into She believes Nathan, who is a fine young man,
the meaning also changes. Therefore, projected fact clauses can all be transposi-
tioned with the primary clauses and thus can form absolute clauses.
When the projected non-finite clause construes an act, it is always not clear
whether the nominal group between the main verb and the non-finite element is
attached to the projecting clause or the projected clause. The method to test the
attribution of the nominal group is the transposition of the non-finite clause and
the primary clause, such as 3-18a. If the nominal group can be transpositioned
together with non-finite element, it can be affirmed that the nominal group is
attached to the non-finite clause, hence forming an absolute clause; if not, it is
attached to the primary clause, hence no absolute clauses formed. This can be
summarized as the participant criterion for identifying absolute clauses, that is, the
second criterion.
Criterion Two: The subject of absolute clauses is not a direct participant of
the primary clauses.
The subject in both sentences in 3-19 consists of a nominal group and a non-
finite element, but the two are different in syntactic structure. In 3-19a, Tom flying
planes badly is an embedded subject clause, and the subject Tom of the embedded
50 3 Approaching Absolute Clauses …
clause is not an participant of the main verb, rather it is independent from the pri-
mary clause, and hence forming an absolute clause. In 3-19b, People living nearby
is a nominal group, the non-finite element living nearby functioning as the post-
modifier of people which functions as the subject of the main verbal group were
told to, hence not forming an absolute clause. The absolute clause in 3-19a func-
tions as the subject of the main verbal group, so Curme (1931) refers to this kind
of absolute clauses as “absolute nominative in subject clauses.” Constructions with
an embedded absolute nominative in subject clause are not clause complexes, but
simple clauses, and therefore, transposition is out of question.
Based on the above analysis, absolute clauses can be returned to their original
finite form, and the subject of absolute clauses is not that of the primary clause. In
3-20, the string the other sucking a mint is also a “noun phrase + non-finite ele-
ment” structure. It fulfills the requirement that absolute clauses have their finite
equivalents and the subject of absolute clauses is not co-referential with that of the
primary clauses. For example,
3-20 One was knitting, and the other was sucking a mint. (COCA_FIC)
Absolute clauses are clauses without Finite in form. The string the other suck-
ing a mint is the same as an absolute clause in form, but it is actually a paratactic
clause with the finite verb was omitted. A characteristic of omission is the con-
formity of syntactic functions (Zhu et al. 2001: 104). Therefore, a finite clause
with the Finite omitted does not form an absolute clause. The Finite element is
also was in the finite equivalents of the absolute clauses in 3-23. However, the was
here is not the omission of structural equatives. It is not co-referential with the ver-
bal group ’d put up in the primary clause. For example,
3-23 He’d put up a couple of mirrors, one in the downstairs hall, one upstairs on
the landing. (CLOB_L)
This indicates that the non-co-referentiality of verbs can be taken as another
criterion for identifying absolute clauses, and we have the third criterion.
Criterion Three: The verbal group of an absolute clause is not co-referen-
tial with that of the primary clause.
3.6 Summary
clauses. Thus, absolute clauses have the same meaning potential as non-finite
clauses, for example, functioning as adjunct, appositive, subject, and comple-
ment. Following the SFL definition of absolute clauses, we developed the criteria
of identifying absolute clauses from relator, participant, and process: (1) absolute
clauses have their finite equivalents introduced by conjunctions; (2) the subject
of absolute clauses is not a direct participant of the primary clauses; and (3) the
verbal group of an absolute clause is not co-referential with that of the primary
clause. Some problems in traditional research of absolute clauses can be resolved
in terms of these three identification criteria. For example, if with in augmented
absolutes is considered as a conjunctive preposition introducing non-finite clauses,
it does not change the syntactic structure of absolute clauses nor their logico-
semantic relations with the primary clauses. However, the explicit conjunctive
expression with results in that absolute clauses are no longer independent in struc-
ture. On the other hand, if with is considered as a preposition, the complement
of the preposition is not a non-finite clause, but a nominal group containing a
post-modifier.
Due to the complexity of syntactic types, the definition and identification
criteria developed in this chapter do not fully reflect all the characteristics of this
type of structure. For example, there are still intermediate stages between the
absolute clauses we identified and non-absolute clauses. This is a reflection of the
notion of cline in SFL. In the fifth chapter, we will discuss in detail the functional
types realized by absolute clauses in terms of the theory of functional syntax and
the independence of different functional types of absolute clauses under the notion
of cline.
References
4.1 Research Questions
4.2 Research Methods
For example, the research of cohesion patterns by Halliday and Hasan (1976) is a
model of combination of qualitative and quantitative research.
Corpus-based qualitative and quantitative linguistics research is becoming a
mainstream paradigm in the study of languages. Qualitative research is the basis
of quantitative research, and quantitative research makes qualitative research more
accurate. Quantitative and qualitative analyses are complementary to each other
(Bunge 1995). In actual studies, qualitative and quantitative methods are always
used with each other in order to accurately qualify on the basis of quantifying.
SFL puts emphasis on the complementarity between qualitative research and
quantitative research in both theoretical investigation and specific practice. The
discussion of function types and the development of criteria of identifying abso-
lute clauses in the framework of SFL should be carried out on the basis of qual-
itative description and quantitative analysis. The qualitative research of absolute
clauses provides a theoretical basis for the quantitative research.
“Quantitative research is a means for testing objective theories by examining
the relationship among variables” (Creswell 1994: 1). The quantitative research in
the framework of SFL is a realization of the notion of probability. Corpus is an
essential tool for carrying out quantitative research. Corpora help to count up the
overall occurrences of absolute clauses, the proportion of each type of function
realized by absolute clauses, and the occurrences of each function type of abso-
lute clauses. It is also helpful for counting up which types of function of absolute
clauses can or cannot be introduced by with and their probabilities of occurring.
Historical corpora can help to analyze the diachronic changes of the probabilities
and to explain the historical evolution of absolute clauses. The change of the abso-
lute clauses introduced by with can be used to test whether the independence of
absolute clauses in clause complexes is increasing or decreasing, and to further
test whether the use of clause complexes is increasing or decreasing. Synchronic
corpora can be helpful to analyze the distributions of absolute clauses in different
stylistic works and further to summarize and explain the stylistic distributions of
absolute clauses.
“SFL research is corpus-based” (Butler 1985: 15). “Two key tools of empirical
linguistics at the turn of the century are the corpus and the computer” (Sampson
2001: 12). “There are, therefore, close ties between corpus linguistics and SFL”
(Neale 2006) and “corpus-based methodology and text-based research have played
a central role in SFL since the beginning” (Matthiessen 2006). Corpus-based
research ties the form and meaning of language closer in certain context, and cor-
pus itself organically links the form, meaning and function of language together.
Halliday and Matthiessen (2004: 34–35) enumerate three plusses relating to the
use of the corpus: First, its data are authentic; second, its data include spoken lan-
guage; third, the corpus makes it possible to study grammar in quantitative terms.
Currently, “SFL is an ‘extravagant’ theory, which consciously provides a rich
description. This helps to explain why SFL corpus-based work is generally slow,
unmechanised and small-scale in comparison with corpus linguistics” (Thompson
and Hunston 2006). The basic situation is that there is “so much theory built over-
head with so little data to support it” (Halliday 1996[2002]). Although there are
56 4 Research Design
4.3 Data Collection
At the second half of the twentieth century, linguistic research began to use intu-
itively invented sentences as the theoretical evidence. This is a reflection of the
subjectivity of the researchers, and hence cannot reveal the actual status of lan-
guage. SFL takes language as a social semiotic and holds that the meaning of
language comes from its use. Following this notion, except for a small number
of directly cited example sentences, all other example sentences used in this
research are actual data retrieved from such corpora as the Brown Family Corpora,
the BNC, COCA, and COHA.
The original purpose of the Brown Family Corpora is to facilitate the comparative
study of British English and American English. The six corpora in the Brown Family
4.3 Data Collection 57
Corpora are the Brown Corpus (1961), the LOB Corpus (1961), the Frown Corpus
(1992), the FLOB Corpus (1991), the Crown Corpus (2009), and the CLOB Corpus
(2009) (Fig. 4.1).
These six corpora cover a span of time of nearly 50 years. The strength of these
six corpora lies in their comparability: The fact that they are constructed according
to the same design, having virtually the same size and the same selection of texts
and registers represented by 500 matching text samples of 2,000 words each, total-
ing 1 million words. This means that we can use the Brown Family Corpora as an
effective tool for tracking the differences between written English in 1961, 1991,
1992, and 2009 (Fig. 4.2; Table 4.1).
We use the BNC and COCA because both corpora have a large vocabulary
(100 million and 460 million words, respectively). This allows enough data for
constructions of relatively lower occurring frequencies. The two corpora cover
a wide range of registers, including not only newspaper, popular magazines,
academic, and fiction, but also spoken. They can facilitate the comparative study
of British English and American English and the stylistic distribution study. They
58 4 Research Design
also provide convenient search tools. The overall occurring frequency of absolute
clauses is relatively low, and different function types of absolute clauses have dif-
ferent occurring frequencies. As for some types of absolute clauses of lower fre-
quencies, such as those with personal pronoun subject, it is not likely to retrieve
enough occurrences to reflect the distribution trend from the Brown Family
Corpora of only 6 million words. However, the occurrences retrieved from so big a
corpus as COCA are enough to reflect the basic distribution of the retrieved item.
COCA has been expanded by 20 million words each year, since the early
1990s. To the June of 2012, COCA has a total number of more than 464 mil-
lion words. Table 4.2 shows that the BNC is 10 % spoken versus 90 % written,
while in COCA, the corpus is nearly evenly divided among such registers as spo-
ken, fiction, popular magazines, newspaper, and academic (20 % in each). For the
research of absolute clauses, the stylistic distribution of absolute clauses can be
counted from the Brown Family Corpora, and their distribution in spoken versus
written registers can be counted from COCA and the BNC (Table 4.3).
Absolute clauses are an ancient syntactic structure. It has been changing with
the evolution of language from Old English through Modern English. Since the
4.3 Data Collection 59
Table 4.3 COCA and BNC compared in terms of register balance and number of words
REGISTER COCA (millions of words) BNC (millions of words)
Spoken 95 10
Fiction 90 17
Popular magazines 96 16
Newspaper 92 11
Academic 91 16
Other 30
Total 464 100
nineteenth century, grammarians have paid enough attention to this structure and
its evolution. Diachronic corpora are an effective tool for the research of the evo-
lution of language. The Brown Family Corpora cover a relatively short span of
time. Although the current application and distribution of absolute clauses can be
acquired, it is not possible to catch the diachronic evolution of absolute clauses in
such a short history as less than 50 years. To address this deficiency, we choose
to use COHA (the Corpus of Historical American English) to conduct diachronic
analyses of absolute clauses. COHA is the largest structured corpus of historical
English with a total number of words of 400 million, covering a span of time from
1810 to 2009. The corpus is balanced by register across the decades. For e xample,
fiction accounts for 48–55 % of the total in each decade (1810s–2000s), and the
corpus is balanced across decades for sub-registers and domains as well (e.g., the
sub-registers of prose, poetry, and drama in the register of fiction). This balance
across registers and sub-registers allows researchers to examine changes of lan-
guage and be reasonably certain that the data reflect actual changes in the “real
world,” rather than just being artifacts of a changing register balance. As for the
research of absolute clauses, this balance helps to examine their changes of fre-
quencies and stylistic distributions during the 200 years of language evolution
(Table 4.4).
60 4 Research Design
4.4 Data Processing
We first classify the six Brown Family Corpora according to the four registers of
press, general prose, learned, and fiction into 24 sub-corpora, tag all the sub-cor-
pora using TreeTagger 2.0 and write the following regular expression according to
the formal types of absolute clauses.
RE1.\S+_(,|SENT|:)\s(\S+_(DT|CD|P\w*)\s)*(\S+_JJ\w*\s)*\S+_(PP|N\w+)\s(\
S+_RB\w*\s)*(\S+_TO\s\S+_V[BDHV]|\S+_(V(B|D|H|V)[GN]|JJ\w*|RB\
w*|IN|DT|N\w+))(\S+_IN\s)*(\S+_(DT|CD|P\w*)\s)*(\S+_(PP|N\w+)\s)*
Then, we use AntConc 3.2.4w for corpus search. The retrieved occurrences
match all the constructions consisting of a nominal group or personal pronoun and
a non-finite element or a verbless element. Finally, we check all the concordance
lines, exclude all the lines that cannot form absolute clauses, and extract all the
absolute clauses. However, we retrieved a total of 95,730 concordance lines using
the above regular expression. It might be a hard job to extract absolute clauses
manually from these nearly 100 thousand concordance lines, and in these lines,
only a very few constructions can form absolute clauses; thereby, it is necessary
4.4 Data Processing 61
to limit the searching conditions to some extent to ensure a relatively higher rate
of absolute clauses in the concordance lines. For example, we can search the con-
structions consisting of a nominal group and a non-finite element directly ending
the structure. Therefore, we write the following regular expressions, respectively,
according to the formal types of absolute clauses.
RE2.\S+_(,|SENT|:)\s(\S+_(DT|CD|P\w*)\s)*(\S+_JJ\w*\s)*\S+_(PP|N\w+)\s(\
S+_RB\w*\s)*\S+_V(B|D|H|V)[GN]\s(\S+_IN\s)*(\S+_(DT|CD|P\w*)\s)*(\
S+_(PP|N\w+)\s)*\S+_(,|SENT|:)
RE3.\S+_(,|SENT|:)\s(\S+_(DT|CD|P\w*)\s)*(\S+_JJ\w*\s)*\S+_(PP|N\w+)\s\
S+_(JJ\w*|RB\w*)\s\S+_(,|SENT|:)
RE4.\S+_(,|SENT|:)\s(\S+_(DT|CD|P\w*)\s)*(\S+_JJ\w*\s)*\S+_(PP|N\w+)\s\
S+_TO\s\S+_V[BDHV]\s(\S+_(DT|CD|P\w*)\s)*(\S+_(PP|N\w+)\s)*\
S+_(,|SENT|:)
RE5.\S+_(,|SENT|:)\s(\S+_(DT|CD|P\w*)\s)*(\S+_JJ\w*\s)*\S+_(PP|N\w+)\s\
S+_IN\s(\S+_(DT|CD|P\w*)\s)*\S+_N\w+\s\S+_(,|SENT|:)
RE6.\S+_(,|SENT|:)\s(\S+_(DT|CD|P\w*)\s)*(\S+_JJ\w*\s)*\S+_(PP|NP\w*)\s(\
S+_(DT|CD|P\w*)\s)\S+_N\w+\s\S+_(,|SENT|:)
The proportion of absolute clauses retrieved from the concordance lines searched
using the restricted regular expressions has substantially increased. We retrieved 817
collocations consisting of a nominal group or pronoun and a present or past partici-
ple from the second regular expression, 1,165 consisting of a nominal group or pro-
noun and an adjective or adverb from the third, 105 consisting of a nominal group
or pronoun and an infinitive from the fourth, 2,703 consisting of a nominal group
or pronoun and a prepositional phrase from the fifth and 417 consisting of a proper
noun or personal pronoun and a nominal group from the sixth. Finally, after exclud-
ing the collocations not being able to form absolute clauses, we got 221 absolute
clauses from the second regular expression, 89 from the third, 3 from the fourth, 66
from the fifth, and 3 from the sixth.
In addition, it being … and there being … are two common types of absolute
clauses. By direct word search, we retrieved 21 collocations of it being … and 29
of there being …, among which there are 6 and 13 absolute clauses, respectively.
Based on the absolute clauses retrieved from the above regular expressions, this
research uses UAM CorpusTool 2.8.12 to carry out comparative analyses on the
distributions of absolute clauses in the Brown Family Corpora, including the dis-
tributions of overall occurrences, region, register and case, etc.
It is convenient to carry out comparative study using the Brown Family
Corpora. However, because of the small number of words and short span of time,
we extracted only a small number of absolute clauses from the Brown Family
Corpora. These occurrences may not be enough to reflect the general character-
istics of absolute clauses. Moreover, these corpora lack the spoken register rep-
resenting the informal style. Therefore, we choose to use the BNC and COHA.
The BNC includes such registers as spoken, newspaper, popular magazines, fic-
tion, academic, and others. The registers of newspaper, popular magazines, fiction,
and academic are roughly equivalent to the classification of registers. However, the
62 4 Research Design
register of popular magazine includes only pop lore in the BNC, while the register
of general prose in the Brown Family Corpora consists of religion, trades, hobbies,
popular lore, belles lettres, biographies, essays and miscellaneous, etc. These con-
tents are included in the registers of miscellaneous and non-academic, and thereby
we choose popular magazine, non-academic, and miscellaneous in the BNC as the
register of general prose. More importantly, the BNC corpus includes the spoken
register representing the informal style.
The BNC is available online, so there is no way to write regular expressions,
or to search using AntConc 3.2.4w. However, we can write the following search
query in terms of the CLAWS7 Tagset.
SQ1. [y*] [nn*]|[p*] [v?n*]|[v?g*]|[nn*]|[j*]|[r*]|[i*]
With this search query, we can retrieve all the collocations consisting of a noun
or pronoun and a present or past participle, noun, adjective, adverb or preposi-
tional phrase following any punctuation. However, the number of concordance
lines is so big that it is quite difficult to extract all the absolute clauses by hand, so
we make appropriate limits on the search query. For example, we can search out
the collocations directly ended with an adjective after a noun or pronoun, or the
collocations directly ended with a prepositional phrase composed of a preposition
and a noun. Therefore, we write the following search queries according to the for-
mal types of absolute clauses.
SQ2. [y*] [nn*]|[p*] [v?g*] (0-2) [y*]
SQ3. [y*] [nn*]|[p*] [v?n*] (0-2) [y*]
SQ4. [y*] [nn*]|[p*] [j*] [y*]
SQ5. [y*] [nn*]|[p*] [i*] [nn*] [y*]
Using the above four search queries, we can retrieve the absolute clauses com-
posed of a noun or pronoun and a present or past participle, adjective and prep-
ositional phrase. The reason why we choose these four types of collocation is
that within all forms of absolute clauses, these four types are of the highest fre-
quency of occurrence and are of a higher probability to form absolute clauses (See
Table 6.1 and Fig. 6.1).
Based on occurrences of absolute clauses retrieved from the above search que-
ries, this research analyzes the stylistic distribution of absolute clauses in the BNC
and carries out a comparative analysis with the research result based on the Brown
Family Corpora.
COHA covers a span of time from 1810 to 2010. We count the absolute clauses
across the decades. Like the BNC, we write the following search query according
to the CLAWS7 Tagset.
SQ1. [y*] [nn*]|[p*] [v?n*]|[v?g*]|[nn*]|[j*]|[r*]|[i*]
COHA has a total number of words of 400 million. It is a hard work to extract
all the absolute clauses from the collocations retrieved from so big a corpus by
hand, so we make some limits on the search query. We only search absolute
4.4 Data Processing 63
References
Neale, A. (2006). Matching corpus data and system networks: Using corpora to modify and
extend the system networks for transitivity in English. In G. Thompson & S. Hunston (Eds.),
System and corpus: Exploring connections (pp. 143–163). London: Equinox.
Quirk, R., Greenbaum, S., Leech, G., & Svartvik, J. (1985). A comprehensive grammar of the
English language. London, New York: Longman.
Sampson, G. (2001). Empirical linguistics. London, New York: Continuum.
Thompson, G., & Hunston, S. (2006). System and corpus: Two traditions with a common
ground. In G. Thompson & S. Hunston (Eds.), System and corpus: Exploring connections
(pp. 1–14). London: Equinox.
Chapter 5
Relationships Realized by Absolute Clauses
5.1 Introduction
SFL attributes absolute clauses to the category of non-finite clauses. Thus, abso-
lute clauses have the same relation potential as non-finite clauses.
In Chap. 2, we discussed the formal and functional types of absolute clauses. The
types of absolute clauses in traditional grammar are represented in Fig. 5.1.
The syntactic analysis of SFL is meaning-based. The meaning-based syntac-
tic analysis facilitates the discussion of covert categories of grammar. Its focus of
research is not on the overt markers. For example, SFL does not distinguish the cases
of nouns or pronouns, taking case as a covert category, but distinguishes such rank
units as clause, group/phrase, word, and morpheme at the lexico-grammatical level.
Absolute clauses are composed of a nominal group and a non-finite verbal group,
nominal group, adverbial group or prepositional phrase at the lexico-grammatical
level. They have the same transitivity structure, thematic structure, and informational
structure as finite clauses at the clause rank. What differs absolute clauses from finite
clauses in structure is the lack of Finite in the mood structure. The SFL-based formal
types of absolute clauses are shown in Fig. 5.2.
noun
NOMINAL
GROUP
pronoun
present
absolute clause participle
nonfinite past
infinitive
VERBAL noun
GROUP
nominal
adverbial adjective
verbless
prepositional
As can be seen, absolute clauses are composed of nominal groups and non-finite
verbal groups in both traditional grammar and SFL. The formal difference between
the two lies in that SFL has no case requirement for the nominal group function-
ing as subject. In addition, SFL does not define the functions realized by absolute
clauses. This opens up a wide potential of functions for absolute clauses.
5.2.2 Relation Potential
According to SFL, absolute clauses are non-finite clauses with an explicit subject
but with no finite or explicit conjunctive expressions. The functions of absolute
clauses are not limited to the traditional clausal adjuncts and attendant circum-
stances. Since absolute clauses are at the clause rank, they have the potential to
realize multiple meanings. First, absolute clauses can enhance or extend the pri-
mary clauses as hypotactic clauses. Hypotactic enhancing clauses are adverbial
clauses in traditional grammar, while attendant circumstances in traditional gram-
mar are hypotactic extension in SFL. When defining lexico-grammatical items,
SFL does not define their functions. Therefore, besides hypotactic expansion,
absolute clauses have the potential to realize other semantic relations. For exam-
ple, absolute clauses can be used independently, and the logico-semantic relations
are realized through textual cohesive devices but not structural devices. Absolute
clauses have also the potential to realize parataxis and hypotactic projection in
clause complexes and to realize participants, or the modifier or appositive of par-
ticipants in simple clauses. This relation potential is represented in Fig. 5.3.
Absolute clauses have the potential to realize all the relationships shown in
Fig. 5.3. In the next section, we will analyze in detail all the relationships that
absolute clauses can actually realize.
68 5 Relationships Realized by Absolute Clauses
parataxis
TAXIS
hypotaxis
idea
secondary projection
locution
LOGICO-
SEMANTIC elaboration
expansion extension
absolute clause independent enhancement
subject
one level
complement
embedded
appositive
two level
modifier
5.3.1 Expansion
5.3.1.1 Elaboration
tones are the same. The difference between hypotactic and paratactic elaborations
is that the former are relative clauses introduced by relative pronouns or relative
adverbs functioning as both structural Theme and topic Theme, which correspond
to the nonrestrictive attributive clauses in traditional grammar, while the latter usu-
ally have no conjunctive expressions realizing elaboration.
Paratactic and hypotactic elaborations are complementary to each other in
meaning. The former involves exposition, exemplification, and clarification and
the latter, description. For example,
5-1a. She wasn’t a show dog; I didn’t buy her as a show dog. (Halliday and
Matthiessen 2004: 398) (exposition)
b. You’re too old for that game; you couldn’t bend over (ibid.: 398)
(exemplification)
c. They weren’t show animals; we just had them as pets. (ibid.: 399)
(clarification)
d. You followed them with The Greenlanders, which seems to me more ambi-
tious. (ibid.: 399) (description)
The two clauses in a paratactic clause complex are both finite. They do not need
conjunctions to realize the relator. If there are conjunctions, they are cohesive rather
than structural markers. For example, in exposition, the secondary clause repeats
more or less the information of the primary clause through lexical repetition or
synonymy to function as emphasis. In exemplification, the secondary clause main-
tains the lexical cohesive relation with the primary clause through hyponymy or
meronymy. In clarification, the secondary clause clarifies the thesis of the primary
clause through synonymy. Hypotactic elaborative clauses are nonrestrictive relative
clauses, a descriptive explanation of a part of or the whole primary clause. The log-
ico-semantic relations between the primary and the secondary clauses are realized
by relative pronouns or relative adverbs. Non-defining relative clauses and defining
relative clauses embedded in nominal groups functioning as qualifier are different
both in meaning and in form. The difference in meaning is that defining relative
clauses define a specific subset, while non-defining relative clauses add new fea-
tures to the already-specific clause or a constituent of a clause. The difference in
form is that the non-defining relative clauses are separated from the primary clauses
with a punctuation mark, while the defining relative clauses usually follow the ante-
cedent immediately.
The semantic and formal differences can be used to answer the following two
questions: (1) Why can the non-subject relatives of defining relative clauses be
omitted? (2) Why cannot the relatives of non-defining relative clauses be that?
The answer to the first question is as follows. When the relative is not the
subject of a relative clause, the subject of the relative clause is not co-referential
with the subject of the clause within which it is embedded. The logico-semantic
relation of the two clauses is clear, so there may be no overt relatives. When the
relative is the subject, the omission of the relative will result in the antecedent
playing a syntactic role in both the relative clause and the main clause. This is
ungrammatical.
70 5 Relationships Realized by Absolute Clauses
The following is the answer to the second question. Since non-defining relative
clauses functions as the explanation but not the restriction of the antecedent, the
relationship between them is relatively loose. If that is used as the subject of the
relative clause, it may be misinterpreted as a demonstrative pronoun functioning as
the subject of a paratactic clause.
The explanations of the two questions show that the rank status of non-defining
relative clauses, that is, the elaborative dependent clauses, is ambiguous. It should
be in between the embedded clauses functioning as modifier and paratactic elabo-
rative clauses, the three forming a cline.
The elaborative secondary clause can also be non-finite. The meaning of non-
finite clauses is usually not as specific as that of their finite equivalents. Their
semantic domain can be a nominal group or a larger syntagma, or even a whole
clause. Elaborative non-finite clauses do not need prepositions or conjunctions to
realize relators, because “relators themselves embody meanings of extension or
enhancement” (Yang 2003: 79). Non-finite dependent clauses may have their finite
equivalents, as in 5-2, or not, as in 5-3.
5-2a. He was an absolute loner of a man, pursuing some dream of exploration in
the jungles. (Halliday and Matthiessen 2004: 403)
b. He was an absolute loner of a man, who pursued some dream of exploration
in the jungles.
5-3 I worked for a local firm at that time, selling office equipment. (ibid.: 404)
Like continuing clauses, the elaborative relation realized by these non-finite
dependent clauses is clarification. In this case, non-finite clauses may have an
explicit subject, as in 5-4a. If the subject is not introduced by a preposition, they
may form absolute clauses, as in 5-4b.
5-4a. It’s a much bigger house, for the children to have their own rooms.
(Halliday and Matthiessen 2004: 403)
b. John went off by himself; the rest of us staying behind. (ibid.: 404)
As is mentioned above, attributive clauses in traditional grammar include nonre-
strictive and restrictive attributive clauses. The antecedent of a restrictive attributive
clause is a nominal group. Attributive clauses function as the modifier of ante-
cedents. Restrictive attributive clauses are embedded clauses in SFL, rankshifted
to function as modifiers in the clauses in which they are embedded. They do not
change the syntactic structure of the clauses in which they are embedded, and the
whole structure is still a simple clause. The antecedent of a nonrestrictive clause can
be a nominal group or the whole primary clause. Nonrestrictive clauses augment the
original clause structure, and the whole construction is a hypotactic clause complex.
5-5a. She opened a bakery, which proved to be an instant success. (COCA_FIC)
b. He had a theory, which was not original. (COCA_FIC)
c. He had a theory, not original.
The non-defining relative clause which proved to be an instant success in
5-5a functions to explain the primary clause She opened a bakery as antecedent. The
5.3 Dependent Absolute Clauses 71
relative clause in 5-5b which was not original is a relational clause, in which not orig-
inal is the attribute of the antecedent a theory. In this case, the finite clause can be
changed into the non-finite form, with the subject and finite carrying no information
omitted, leaving only the attribute carrying information, as in 5-5c. If the attribute
itself is composed of a nominal group and a non-finite verbal group forming a non-
finite clause with subject, this construction forms an absolute clause. For example,
5-6a. I just wondered why I heard a noise, the gerbils running. (BNC_SPOK)
b. The noise doesn’t stop—glass breaking, wood popping, everything smash-
ing down. (COHA_FIC)
c. Then, he was aware of another presence, a kigrin circling him, wading
through the murky water. (COHA_FIC)
Each of the absolute clauses in 5-6 explains a nominal group in the primary
clause, realizing elaboration.
5.3.1.2 Extension
In a clause complex, extension means that “one clause extends the meaning of
another by adding something new to it” (Halliday and Matthiessen 2004: 405).
Three subtypes of paratactic extension can be recognized: addition, variation, and
alternation. Addition can further be divided into positive addition (and), negative
addition (nor) and adversative (but), and variation can be divided into replacive
(instead) and subtractive (except). Hypotactic extension also embraces addition,
variation, and alternation, but the secondary clauses realizing extension are
not continuing clauses but dependent clauses. There is not a sharp line between
paratactic and hypotactic extensions. “If the extending clause could precede
(thereby becoming thematic in the clause complex), the relationship is hypotac-
tic” (Halliday and Matthiessen 2004: 408). These extensive relations are shown in
Fig. 5.4.
Non-finite extending clauses cover the subtypes of addition and variation, but
not alternation, and there is no negative addition in the subtype of addition. Thus,
positive
extension
addition negative
adversative
LOGICO-
alternation
SEMANTIC
replacive
variation
subtractive
72 5 Relationships Realized by Absolute Clauses
positive
extension
addition ----
adversative
LOGICO-
----
SEMANTIC
replacive
variation
subtractive
positive
extension
addition ----
adversative
LOGICO- ----
SEMANTIC
----
At the same time, additive absolute clauses can be introduced by with or with-
out, forming augmented absolute clauses, implying “an accompanying circum-
stance to the situation described in the matrix clause” (Quirk et al. 1985: 1124).
For example,
5-11a. He leaned forward, (with) his hands on his knees. (COCA_FIC) (addition:
positive)
b. They had been working for 18 months, spending lots of money (with)no
income coming in. (COCA_MAG) (addition: adversative)
5.3.1.3 Enhancement
“In enhancement one clause (or sub-complex) enhances the meaning of another
by qualifying it in one of a number of possible ways: by reference to time, place,
manner, cause or condition.” (Halliday and Matthiessen 2004: 410) The two com-
monest ways of enhancing are time and cause. Paratactic enhancement is a kind
of coordination with a circumstantial feature realized by conjunctions or conjunc-
tional groups. Hypotactic enhancing clauses are known in traditional grammar as
adverbial clauses. Finite clauses of enhancement always need conjunctions to real-
ize the dependent relationship and the circumstantial relationship. Such conjunc-
tions are referred to as binders. For example,
5-12a. By the time I got outta that house, my mouth run dry and my knees wob-
bled. (COCA_FIC) (time)
b. House prices will tend to rise where the sun shines. (COHA_MAG) (place)
c. As it happens, railroad shopping is a growth industry in Zurich. (COCA_
NEWS) (manner)
d. We are reducing the numbers of people who qualify for free food because
people are cheating. (COCA_NEWS) (cause)
e. If I had a look, I’d be sure to find it. (BNC_MISC) (condition)
74 5 Relationships Realized by Absolute Clauses
Non-finite clauses of enhancement can also realize the relationships of time, man-
ner, cause, condition, and so on. Such conjunctive expressions as conjunctions,
prepositions, and zero conjunctions can all function as relators. For example,
5-13a. When returning home, the group was instructed to remain in silence and
(COCA_ACAD)
b. On returning home he sets to work. (COCA_MAG)
c. Returning home, Mick throws a party for the local poor. (COCA_FIC)
Non-finite clauses of time, manner, cause, or condition may have explicit con-
junctive expressions. For example,
5-14a. Create an extensive pros-and-cons list before making a decision.
(COCA_MAG) (time)
b. Ford hopes to beat Corker by making the race a referendum on the
Republican Congress. (COCA_NEWS) (manner)
c. I am no parenting expert, despite having four children. (COCA_MAG)
(concession)
d. But in real life, most people don’t eat fresh produce without doing some-
thing to it first. (COCA_MAG) (condition)
Non-finite clauses of time and cause (reason, purpose, and result) may have no
explicit conjunctive expressions. For example,
5-15a. Having finished her little speech, she turned gracefully and left the room.
(COCA_FIC) (time)
b. Being a Stone elemental, I could feel the vibrations in the rock
(COCA_FIC) (cause: reason)
c. To win the war, early detection of the deep and hidden cancers will be key.
(COCA_MAG) (cause: purpose)
d. The fish circled four times, then dove from sight, never to be seen again.
(COCA_FIC) (cause: result)
Like non-finite clauses of extension, the logical subject of non-finite clauses
of enhancement without explicit subject is co-referential with the subject of the
primary clauses. If the subject of non-finite clauses is not co-referential with
the subject of the primary clause, the subject of the dependent clause is always
the speaker, or the agent of the receptive clause, or even some certain unimpor-
tant nonspecific entity. However, such dependent clauses usually have their own
explicit subject. When this explicit subject is case marked, it may be oblique (e.g.,
him) or possessive (e.g., his). For example,
5-16a. In order for him to salvage his legacy, he’s going to have to really do what
he can to maintain some goodwill. (COCA_SPOK)
b. With him being down, everybody on the other team piled on him.
(COCA_FIC)
c. With his being on the road so much, he wanted Dusty to be able to defend
herself and her mother. (COCA_FIC)
5.3 Dependent Absolute Clauses 75
5.3.2 Projection
Absolute clauses as non-finite clauses can only realize hypotaxis. SFL distin-
guishes two types of hypotaxis: hypotactic expansion and hypotactic projection.
Now that absolute clauses can realize hypotactic expansion, then can they real-
ize hypotactic projection? According to the criteria of identifying absolute clauses
developed in Chap. 3, non-finite clauses of expansion can form expanding abso-
lute clauses, and non-finite clauses of projection can also form projecting absolute
clauses.
Traditional grammar labels absolute clauses with adjuncts. However, Curme
(1931: 157) identifies two classes of nominative absolutes according to func-
tion: absolute nominative in adverbial clauses and absolute nominative in subject
clauses. The subject of the absolute nominative in subject clauses is not structur-
ally related to the main verb of the main clause. For example,
5-20a. Three such rascals hanged in one day is good work for society. (Curme
1931: 157)
b. Women having the vote reduces man’s political power. (ibid)
c. Them trying to sing a song was just too horrible. (Reuland 1983: 101)
Seen from the transitivity structure, clauses functioning as both subject and
complement realize participants. Since clauses of subject can form absolute
clauses, clauses of complement can also form absolute clauses. Accordingly, this
kind of absolute clauses functioning as complement can be referred to as absolute
nominative in compliment clauses. For example,
5-21 We heard the gun firing several times. (BNC_FIC)
However, this view is in conflict with that of the traditional grammar, accord-
ing to which the gun in 5-21 functions as the object of heard and firing several
times as the complement of the object. According to SFL, the non-finite clause the
gun firing several times as a whole functions as the complement of heard. This
means that “functional syntactic analysis can better realize the meaning of clauses
than traditional syntactic analysis” (Huang 1998b), because what we heard is the
sound of the gun, but not the gun itself. In fact, the gun firing several times in 5-21
is a subject–predicate construction. The subject the gun is not directly related to
the verbal group heard, and there is no explicit conjunction (complementizer) that
between them. If the subject clause in 5-20 is an absolute clause, then the comple-
ment clause in 5-21 can also be seen as an absolute clause. It is certain that not all
“nominal group + participial phrase” constructions functioning as subject or com-
plement form absolute clauses. This is because this kind of word strings may have
two different interpretations: (1) The participial phrase functions as the post-mod-
ifier of the nominal group; and (2) the participial phrase functions as the logical
predicate of the nominal group. In the case of the former, the whole construction is
a nominal group, the noun being the head of the nominal group functioning as the
subject of the main verb, and the non-finite element being an embedded element
5.3 Dependent Absolute Clauses 77
functioning as post-modifier. For example, in 5-22a, the subject of the main verb
is Coracles, and the non-finite element using nets functions as post-modifier. In
the latter case, however, the noun and the non-finite element together function as
the subject of the main verb, forming an absolute clause, as Children having fun in
5-22b.
5-22a. Coracles using nets were banned from the Wye in the twenties. (BNC_NEWS)
b. Children having fun is quite a structured exercise. (BNC_MAG)
In SFL, the subject clause here is embedding, the whole construction being
a simple clause, but the complement clause is projection, the whole construc-
tion being a clause complex. In discussing the relationships of clause complexes,
Halliday and Matthiessen (2004: 443) define three projection systems: (i) the level
of projection (idea versus locution), (ii) the mode of projection (hypotactic report-
ing versus paratactic quoting), and (iii) the speech function (projected proposition
versus projected proposal). Thus, the system of projection is shown in Fig. 5.7.
5-23a. Ben said, “Why are you dressed like that?” (COCA_FIC)
b. Ben said he was working on the ninth of the twelve steps. (COCA_FIC)
c. Kemp said, “Don’t put ice in that.” (BNC_FIC)
d. I told you not to disturb me. (BNC_FIC)
e. Mller said, “Goodbye, Madame.” (BNC_FIC)
In 5-23a, the level of projection is locution, the mode of projection is paratactic
quoting, and the speech function is proposition (statement). In 5-23b, the level of
projection is idea, the mode of projection is hypotactic reporting, and the speech
function is proposition (statement). In 5-23c, the level of projection is locution, the
mode of projection is paratactic quoting, and the speech function is proposal. In
5-24d, the level of projection is idea, the mode of projection is hypotactic reporting,
locution
LEVEL
idea
reporting
MODE
projection
quoting
statement
proposition
major question
FUNCTION proposal
minor
Fig. 5.7 System of projection
78 5 Relationships Realized by Absolute Clauses
and the speech function is proposal. In addition to proposition and proposal, the
paratactic projected elements can also be minor speech functions.
Paratactic projected clauses can be used independently, without any signs
of being projected. In this sense, paratactic projected non-finite clauses have the
potential to form absolute clauses. However, non-finite clauses tend to realize
hypotaxis rather than parataxis and hence cannot form absolute clauses. The hypo-
tactic projected clauses of proposition are usually finite clauses, as in 5-24a. There
are only a small number of “accusative + infinitive” constructions that can be non-
finite clauses, as in 5-24b. The hypotactic projected clauses of proposal can be
finite ones, as in 5-24c, or non-finite ones, as in 5-24d.
5-24a. Martha thought that the river was angry. (BNC_FIC)
b. I believe it to be the source of evil. (BNC_FIC)
c. Lightman ordered that Kruger be publicly reprimanded on two counts.
(BNC_NEWS)
d. He told me to leave my job. (BNC_FIC)
As discussed in Chap. 2, traditional grammar defines absolute clauses in both
form and function. SFL considers language as meaning potential, emphasizing
using language to create meaning. It does not restrict the functions being able to
be realized by a language form. Both non-finite clauses in 5-24b and 5-24d have
their own subject. It is certain that not all non-finite clauses with subject can nec-
essarily form absolute clauses. The subject of the non-finite clause in 5-24b is it.
In traditional grammar, it is the object of believe, and to be the source of evil is the
complement of it. In SFL, the non-finite clause it to be the source of evil functions
as the complement of believe, and the non-finite clause itself has its own func-
tional syntactical structure.
I believe it to be the source of
evil
TG analysis Subject Predicate Object Complement
SFL analysis Subject Finite Main Complement
verb Subject Predicate
Transitivity Senser Process: mental Identified Process: relational Identifier
structure Metaphenomenon
Logico-semantic Primary Secondary
Interdependency α ‘β
The non-finite clause in both 5-24b and 5-19d can be considered as an ele-
ment of the primary clause. Both have their own functional syntactic structure
and both are independent from the primary clause. The only difference is that the
non-finite clause in 5-24b realizes participant and that in 5-19d realizes circum-
stance. According to traditional grammar, 5-24b is a simple clause. Although the
object and complement form a subject–predicate construction, they cannot form
an absolute clause, because the object is controlled by the main verb. 5-19d is a
clause complex, and the subject of the logical subject–predicate construction is
not controlled by the main verb, and hence, they can form an absolute clause. The
secondary clauses of both clause complexes of expansion and of projection distin-
guished by SFL can be non-finite clauses. The adverbial clauses distinguished by
traditional grammar can be non-finite clauses, but the object clauses have no finite
equivalents. Because of this, absolute clauses defined by traditional grammar only
serve to function as adjuncts. According to SFL, the projected clauses functioning
as complement have both finite and nonfinite forms. Since absolute clauses can
realize circumstances, they should also be able to realize participants.
Weather permitting, they shuttle up and down
If weather permits, in packs
Transitivity Relator Actor Process: Actor Process: Circumstance
structure material material
Circumstance
Logico-semantic Secondary Primary
Interdependency ×β α
I believe it to be the source
Transitivity Senser Process: (that) it is of evil
structure mental Identified Process: Identifier
relational
Metaphenomenon
Logico-semantic Primary Secondary
Interdependency α ‘β
He told me to leave my
(that) I should leave job
Transitivity Sayer Process: Recipient Actor Process: Goal
structure verbal material
verbiage
Logico-semantic Primary Secondary
Interdependency α ‘β
The reported proposal can be locution or idea. For example, 5-24d is a reported
locution. The subject of a reported locution is implicit and it requires the recipi-
ent’s reasoning. For example,
5-25a. She had promised him to obey his commands. (COHA_FIC)
b. *He had been promised she to obey his commands.1
c. *He had been promised by her to obey his commands.
d. He had been promised that she would obey his commands.
5-26a. I promised him a wheelbarrow to be pushed every day in the resolution of
his debt. (COHA_FIC)
b. He was promised a wheelbarrow to be pushed every day in the resolution
of his debt.
Based on the above analysis, the recipient me in 5-24d is the subject of the pro-
jected non-finite clause, and the projecting clause can be changed to the passive
form I was told to leave my job. The subject I of the primary clause is still the
subject of the non-finite clause. The recipient him in 5-25 belongs to the primary
clause in structure, and it is not the subject of the non-finite clause, and therefore,
they do not form an absolute clause.
She had him to obey his
promised (that) would obey commands
she
Syntactic Subject Predicate Complement 1 Subject Predicate Complement
structure Complement 2
Transitivity Sayer Process: Recipient Actor Process: Goal
structure verbal material
Verbiage
Logico-semantic Primary Secondary
Interdependency α ‘β
The subject of a passive reporting locution is explicit, but the Actor is still implicit.
Since the subject of a passive clause is not constrained by the main verb of the project-
ing clause, passive non-finite clauses can form absolute clauses. See example 5-26:
I promised him a wheelbarrow to be pushed …
(that) a be pushed
wheelbarrow
1 * Unacceptable
5.3 Dependent Absolute Clauses 81
As can be seen, the Sayer she in 5-25a is the implicit subject of the projected
non-finite clause. If the projecting clause is changed into passive voice, the subject
of the primary clause is not the subject of the non-finite clause. In other words, the
subject of the non-finite clause is not co-referential with that of the primary clause.
Therefore, the subject of the non-finite clause should be made explicit. There are two
ways: First, changing the non-finite clause into an absolute clause; second, using
preposition by to introduce the explicit Agent. Following the first way, we get 5-25b.
According to Halliday and Matthiessen (2004), 5-25b is grammatically unac-
ceptable. This is because although she is the explicit subject of the non-finite
clause, it is also the Agent of promised. In a passive clause, the Agent and the
subject do not overlap, so the subject should not be in the nominative case. If the
subject of the non-finite clause is not the Agent of the main verb of the primary
clause, this construction is grammatically acceptable. For example,
5-27a. Wright was promised an opportunity to advance. (COHA_MAG)
b. ?To advance, Wright was promised an opportunity.2
5-28a. For some time we have been promised legislation to reform the laws on
friendly societies. (BNC_MISC)
b. To reform the laws on friendly societies, for some time we have been
promised legislation.
5-29a. The first 200 attendees were promised passes to the Tom Hanks film.
(COHA_NEWS)
b. *To the Tom Hanks film, the first 200 attendees were promised passes.
However, although 5-27a, 5-28a, and 5-29a are similar to 5-25b in form, the two
structures are different in syntactic function. The subject of the nonfinite clause to
advance in 5-27a is co-referential with that of the primary clause. However, syntacti-
cally to advance is embedded in the nominal group an opportunity functioning as its
post-modifier. The two elements together are projected by the main verb promised. In
other words, the projected element is not a macrophenomenon, but a phenomenon, and
the whole construction is a simple clause rather than a clause complex and hence can-
not form an absolute clause. This can be better demonstrated in 5-29a. 5-28a can also
be analyzed like this, except that the subject of the non-finite clause in 5-28a is not co-
referential with that of the primary clause. The subject of the non-finite clause to obey
his commands in 5-25b is not co-referential with that of the primary clause; to obey his
commands is the projected element of promised, and it is a macrophenomenon.
2 ? Maybe acceptable
82 5 Relationships Realized by Absolute Clauses
Following the second way, we get 5-25c. In 25c, the initiator her of the verb
promised is introduced by the preposition by, but by cannot at the same time intro-
duce the subject of the nonfinite clause. That is to say, the subject of the non-finite
clause remains implicit.
Therefore, the only way to make the subject of the non-finite clause explicit is
to change the non-finite clause into finite, as in 5-25d. From another perspective, if
the subject of the reported locution is explicit, it does not need the recipient’s rea-
soning. There will be a different situation. For example,
5-30 They promised her that the reign of terror would be over before the year
was out. (BNC_FIC)
5-31a. Russia was promised that NATO would not expand. (COCA_ACAD)
b. ?Russia was promised NATO not to expand.
In 5-26a, a wheelbarrow is the subject of the projected non-finite clause. It is
not directly controlled by the main verb. Therefore, the projected passive clause a
wheelbarrow to be pushed every day in the resolution of his debt may be consid-
ered as an absolute clause. The passive form of the projecting clause can still pro-
ject a non-finite clause, as in 5-26b and 5-32, because the subject of this projected
nonfinite clause is originally explicit.
5-32 He was offered counselling to be paid for by the church. (BNC_K1S)
Another interpretation of the infinitives in 5-27a and 5-28a is that they are expand-
ing clauses of purpose. For example, the two clauses in a hypotactic expanding
clause complex can be translocationed. This can be better explained in 5-33.
5-33a. He also promised measures to overcome two obstacles…. (BNC_NA)
b. He also promised measures were to be taken.
c. He also promised measures to be taken.
d. He also promised measures to overcome two obstacles were to be taken.
e. He also promised measures were to be taken to overcome two obstacles.
f. To overcome two obstacles, he also promised measures were to be taken.
In 5-33a, the string to overcome two obstacles functions as the post-modifier of the
projected element measures or as an expanding non-finite clause of purpose. In 5-33b,
the projected finite clause measures were to be taken can be changed into the non-finite
clause in 5-33c, forming an absolute clause. This is because measures and to be taken
form a non-finite clause with subject. It is obvious that to overcome two obstacles in
5-33d functions as post-modifier and that in 5-33e is an expanding non-finite clause of
purpose because it is not immediately after the head noun measures that it modifies, and
moreover, it can be translocated with the primary clause, as is shown in 5-33f.
We can give further explanations to absolute clauses of reported locution in
5-34 and 5-35.
5-34 The King held it good to accomplish her desire; and forthwith ordered let-
ters to be drawn up to Rodrigo of Bivar. (BNC_NA)
5-35 He ordered fighters to shoot down Hassans Boeing. (BNC_MISC)
5.3 Dependent Absolute Clauses 83
In 5-34, letters is not the object of ordered; rather, it is the subject of the non-
finite to be drawn up to Rodrigo of Bivar. They together form an absolute clause.
In 5-35, fighters is the object of ordered. It functions as the subject of the non-
finite clause to shoot down Hassans Boeing, but they do not form an absolute
clause because the non-finite clause to shoot down Hassans Boeing cannot be
translocated with the primary clause to realize the relationship of purpose.
Different from reported locution, the subject of a reported idea is explicit and
it is part of the projected proposal in structure. That is why the projecting clauses
cannot be changed into passive voice, but the projected clauses can. For example,
5-36a. He wanted me to design the top-floor atelier. (COCA_MAG)
b. He wanted the top-floor atelier to be designed by me.
According to traditional grammar, him in 5-25a, me in 5-36a, and the top-floor
atelier in 5-36b are all controlled by the main verb, they function as the object of
the main verb, and the infinitive to obey his commands in 5-25b and to be designed
by me in 5-36b both function as the complement of the object. According to SFL,
however, in 5-25a, him is the complement of the main verb and to obey his com-
mands is the projected clause, and me in 5-36a and the top-floor atelier in 5-36b
are included in the non-finite clauses but not controlled by the main verb in struc-
ture. Therefore, this kind of non-finite clauses with subject consisting of a nominal
group or a pronoun and a non-finite element can form absolute clauses, realizing
macrophenomena in transitivity structure.
He wanted me to design the top-floor atelier
Syntactic structure Subject Predicate Subject Predicate Complement
Complement
Transitivity Senser Process: Actor Process: Goal
structure mental material
Phenomenon
Logico-semantic Primary Secondary
Interdependency α ‘β
Not only can two clauses constitute a clause complex realizing the logico-semantic
relations of parataxis and hypotaxis, but also one of the two clauses can be rank-
shifted to embed in the other clause serving as a constituent of this clause or
embed in a constituent of the clause serving as the modifier of this constituent.
“There is no direct relationship between an embedded clause and the clause within
which it is embedded.” (Halliday and Matthiessen 2004: 426) Therefore, a clause
containing an embedded clause is still a simple clause but not a clause complex.
According to SFL, clauses are composed of groups or phrases, and a clause can
be rankshifted into a group to function as a constituent of another clause or can be
84 5 Relationships Realized by Absolute Clauses
The core element of a non-finite clause is the non-finite verb. Traditional gram-
mar identifies three types of non-finite verbs: participles, infinitives, and gerunds.
Participles can be further divided into present participles and past participles.
Present participles and gerunds both are the -ing form of verbs, and sometimes,
it is hard to distinguish between them. Traditional grammar also categorizes non-
finite verbs into -ing verbs, -ed verbs, and infinitives. The -ing verbs include pre-
sent participles and gerunds. The two ways of classification of non-finite verbs are
shown in Figs. 5.8 and 5.9.
The first classification tends to be function-oriented and the second, form-
oriented. Although the second classification has been universally accepted by
grammarians nowadays, the first classification is obviously briefer and clearer.
5.4 Embedded Absolute Clauses 85
Fig. 5.8 Classification of infinitive
non-finite verbs (I)
present
nonfinite verb participle
past
gerund
Fig. 5.9 Classification of infinitive
non-finite verbs (II)
present participle
nonfinite verb -ing
gerund
-ed
Infinitives and gerunds can both function as subjects or objects, and infinitives and
participles (present participles and past participles) can both function as adjuncts.
The syntactic functions of infinitives equal to the combination of those of gerunds
and participles.
SFL considers the traditional adjectival clauses and the nominal fact clauses as
embedded elements. As for non-finite clauses, SFL adopts the classification shown
in Fig. 5.9. It does not distinguish present participles and gerunds. For example,
Yang (2003: 52–54) distinguishes four types of constructions that non-finite -ing
and to- clauses as embedding can occur, they are subject, complement, modifier,
and the so-called anticipatory subject of cleft sentences. Non-finite -ed clauses
occur in two constructions, in which the non-finite -ed clause is embedded as mod-
ifier of the preceding element or as the complement of the clause. Although SFL
does not distinguish between gerunds and present participles, from these syntactic
functions, it can be seen that the non-finite participial (-ing and -ed) complement
functions as Attribute in relational clauses, while the gerundial complement real-
izes participant. For example,
5-39a. Not being fluent in English results in a language barrier that negatively
affects the patient’s educational process. (COCA_ACAD)
b. They don’t mind getting their hands dirty. (COCA_ACAD)
c. The door remains closed a minute and then there is a soft knock.
(COHA_FIC)
Non-finite elements in 5-39a and 5-39b act as subject and complement, respec-
tively, realizing participants. The former is the Actor of the material clause and the
latter, the Phenomenon of the mental clause. The non-finite element locked in 5-39c
functions as complement, realizing the Attribute of the relational verb remains.
The traditional appositive clauses “seem to have the hypotactic relation of elab-
oration, but the non-finite clauses are only used to describe groups in the finite
clauses” (Yang 2003: 53). Although appositive clauses and attributive clauses
86 5 Relationships Realized by Absolute Clauses
are the same in form, they are different in terms of rankshift. Appositive clauses
belong to the category of nominal clauses, the antecedents of which are usually
general nouns, such as the fact and the idea. Attributive clauses belong to the cat-
egory of adjectival clauses. According to SFL, appositive clauses realizing partici-
pants are at the group rank, while attributive clauses realizing modifiers are at the
word rank. As can be seen, although clauses realizing participants and those realiz-
ing modifiers are both embedded elements, they are different in embedding depth.
In the following section, we will distinguish types of embedded clauses from the
perspective of the embedding depth.
5.4.2 Types of Embedding
c. They believe that the minimum wage could threaten their jobs. (ibid.: 193)
d. I have no idea when he will come back home. (ibid.: 193)
e. We have 30 men who are working from 6 am to 11 pm. (ibid.: 195)
f. And she said that everything was mouldy. (ibid.: 196)
It seems that the dependent clauses in clause complexes can all be taken as embed-
ded clauses, and therefore, hypotactic clause complexes can all be reanalyzed as sim-
ple clauses. The dependent clauses are rankshifted and embedded into the dominate
clauses. Only paratactic clauses cannot be seen as embedded elements. The syntactic
functions of the secondary clauses in 5-41 are shown in the following figures.
5.41a Maya is drinking her bourbon tonight because Vern left today
Syntactic Subject Predicate Complement Adjunct Subject Predicate Adjunct
structure Adjunct
Transitivity Actor Process: Goal Circumstance Relator Actor Process: Circumstance
structure material material
Circumstance
5-41c They believe that the minimum wage could threaten their jobs
Syntactic Subject Predicate Subject Predicate Complement
structure Complement
Transitivity Senser Process: Relator Actor Process: material Goal
structure mental Phenomenon
5-42e Mr. Fedders had borrowed heavily since joining the S.E.C.
Syntactic Subject Finite Predicate Complement Predicate Complement
structure Adjunct
Transitivity Carrier Process: relational Attribute Relator Process: Goal
structure material
Circumstance
5-43a. Birds which are sold into the pet trade are notoriously badly treated.
(BNC_NA)
b. The book which he wrote was circulated to the English universities.
(BNC_MISC)
c. The days when he could do it, or would, were now over. (BNC_FIC)
The antecedents Birds, the book, and the days in 5-43 function as subject, com-
plement, and adjunct, respectively. When the antecedent functions as the subject of
the embedded clause, the embedded clause can be non-finite, as in 5-44, but when
the antecedent functions as complement or adjunct, there are no non-finite forms.
Therefore, non-finite clauses functioning as post-modifier cannot form absolute
clauses.
5-44 People standing nearby laughed softly in derision. (COCA_FIC)
Appositive clauses are relatively special. On the one hand, they belong to the
category of nominal clauses, and on the other hand, they are not the direct partici-
pants of the main verbs. They are embedded in the nominal groups functioning as
participants of the clauses. For example,
5-45a. I thought that he was only a captain. (COCA_FIC)
b. We all regretted that he was left behind. (COCA_FIC)
c. I regretted the fact that I was late. (COCA_FIC)
The project clause in 5-45a is an idea and that in 5-45b is a fact. According
to the foregoing analysis, the former is one-level embedding and the latter, two-
level embedding. It is not directly projected by the projecting clause Mark Antony
regretted; rather, it is the embedded element of the phenomenon the fact, as in
5-45c. The difference between appositive clauses and modifier clauses in form
lies in that the antecedent of the former does not assume an element of the clause,
while that in the latter, assumes an element. Therefore, appositive non-finite
clauses with subject have the potential to form absolute clauses. For example,
5-46 ?The fact, Caesar being ambitious, worries me so much.
Appositive clauses are between modifier clauses and complement clauses in
rank. Modifier clauses have no structural relations with the clauses in which the
nominal groups within which they are embedded assume functions. Subject and
complement clauses both assume syntactic functions in the clauses within which
they are embedded. Appositive clauses are intermediate between them two; they
may assume syntactic functions or not. For example,
5-47a. I can’t enjoy your favorite exercise, swimming in winter. (Zhang 1997: 1355)
b. He had one aim—to fight for the independence of his motherland. (ibid)
In 5-47a, swimming in winter is the appositive of your favorite exercise. It
can also realize participants, that is, they two have the same syntactic function
potential. In 5-47b, to fight for the independence of his motherland is the apposi-
tive of one aim, but they do not realize participants, that is, they are syntactically
5.4 Embedded Absolute Clauses 91
5-49a. Then she said, “I’ll see you in five weeks.” (BNC_FIC)
b. It was stood close beside her in the grass, and said, “Hello, Peggie.”
(BNC_MISC)
However, paratactic projection is different from paratactic expansion at least in
the following four aspects. First, the two clauses in a paratactic clause complex of
expansion can be linked by the conjunction and, but the two clauses in a paratactic
clause complex of projection cannot. Second, both the two clauses in a paratactic
clause complex of expansion can be independently used, but the primary clause in
a paratactic clause complex of projection cannot. Third, the two clauses in a para-
tactic clause complex of expansion cannot be transpositioned, and even if they can,
the one in front position is always the primary clause, but the two clauses in a para-
tactic clause complex of projection can be transpositioned, and the status of each of
the two clauses will not change. Fourth, the two clauses in a paratactic clause com-
plex of expansion are at the same rank, but in a paratactic clause complex of projec-
tion, the projected clause can be a finite one as well as other constructions, such as
minor clauses. However, minor clauses are not clauses in the strict sense, and they
are at the same rank as the projecting clauses. Therefore, the parataxis of projec-
tion is actually not real parataxis. They share many features of hypotaxis and hence
belong to a special parataxis between parataxis and hypotaxis in the real sense.
From the above analysis, parataxis of projection can also be considered as
embedding, but not deep embedding. From the perspective of embedding depth,
all embedding relations can form a cline (see Fig. 5.11).
Embedded clauses defined by SFL include embedded defining clauses, apposi-
tive clauses, and subject clauses, but its definition on complement clauses is rela-
tively unclear. For example, Halliday and Matthiessen (2004) refer to complement
clauses as projected clauses, but some projected non-finite clauses as embed-
ded clauses. In fact, Fig. 5.11 shows that the dependency of absolute clauses is
changing gradually with that of the depth of embedding; there are no absolutely
independent absolute clauses. These above seven types of structure are shown sys-
tematically in Fig. 5.12.
When one of the two clause in the “clause + clause” structure is non-finite, this
relation system will greatly change, as shown in Fig. 5.13.
As can be seen, the rank status of non-finite clauses is lower than that of finite
clauses. This is mainly shown in that non-finite clauses cannot realize paratac-
tic expansion, projected non-finite clauses are characterized with hypotaxis,
and hypotactic non-finite clauses of projection are treated as simple clauses by
Halliday and Matthiessen (2004), etc.
From the above analysis, we find that except for the clause complexes of para-
tactic expansion, other relation types of clause complexes and simple clauses with
clausal subject, appositive, and modifier can be regarded as embedding. They
embed in different depths. In these types of embedded clauses, only the embed-
ded clauses of modifier cannot form absolute clauses because the antecedents of
the embedded clauses always assume certain syntactic functions in the embedded
clauses. Other types of embedded clauses can all form absolute clauses.
5.4 Embedded Absolute Clauses 93
Adjunct clause
Complement clause
Subject clause
Appositive clause
Modifier clause
adjunct
clause +clause hypotaxis
complement
subject
embedding appositive
modifier
adjunct
clause + nonfiniteclause hypotaxis
complement
subject
embedding appositive
modifier
94 5 Relationships Realized by Absolute Clauses
clause being the initiating clause and the secondary clause being the continuing
clause, and the logico-semantic relation between the two clauses is not elaboration
but extension. There are two reasons. First, the two clauses cannot be translocated.
For example, 5-51a can be rewritten as 5-51b. Second, in a hypotactic clause com-
plex, only the primary (dominant) clause is free, and the secondary (dependent)
clause cannot occur independently. However, clauses like which surprised every-
one are able to be used independently.
5-52a.
But to buy some more you need more cash. Which you haven’t got.
(BNC_FIC)
b.
It is therefore upward rankshift; which we must consider. (Halliday
1966[2002]: 121)
Even if the logico-semantic relationship between the two clauses is not exten-
sion but elaboration, the clause introduced by which can still only be considered
realizing parataxis. For example,
5-53a. He was at the zoo, which had just been given a rare species of bird.
(COCA_FIC)
b. *He was at the zoo, and it had just been given a rare species of bird.
The two clauses in 5-53a cannot be translocated and hence is paratactic in
Huang’s (1998a, b, c) terms, and the continuing clause which had just been given
a rare species of bird functions as the explanation of the zoo, so the logico-seman-
tic relation between the two clauses is elaboration. 5-53b is unacceptable because
the conjunction and realizes extension. In this sense, the hypotactic clause com-
plexes consisting of two finite clauses discussed by Halliday (1994: 227–229) and
Halliday and Matthiessen (2004: 400–403) actually realize parataxis, and the log-
ico-semantic relation is elaboration or extension (Huang 1998a, b, c).
Instances 5-1d and 5-53a in traditional grammar are both nonrestrictive attribu-
tive clauses. The difference between the two is that the antecedent of the second-
ary clause of the former is the whole primary clause; it adds information to the
primary clause, realizing extension. The antecedent of the secondary clause of the
latter is a nominal clause in the primary clause; it is the explanation of the nominal
clause, realizing elaboration. The absolute clause the gerbils running in 5-6 and
those in 5-54 cannot be translocated with the primary clauses and hence are abso-
lute clauses of elaboration. This can also be used to explain why these absolute
clauses cannot be introduced by with. Then, can these absolute clauses be consid-
ered realizing parataxis?
5-54a. She could hear voices ahead, a girl squealing, a man laughing.
(BNC_FIC)
b. The pose is natural—relaxed—the legs slightly apart, the arms hanging
down. (COCA_FIC)
These absolute clauses are all relational clauses with Subject and Finite o mitted,
and the residues the gerbils running, a girl squealing, and a man laughing, etc.,
96 5 Relationships Realized by Absolute Clauses
themselves form absolute clauses. At the same time, absolute clauses can be used
independently. For example,
5-55 There was a lot of background noise. People chatting, laughing. (BNC_FIC)
Although the absolute clause in 5-55 elaborates the preceding clause semantically,
the two clauses do not form a clause complex. They are two independent clauses
realizing parataxis. The question is can the translocatability and independability
be the necessary and sufficient conditions for the two clauses to realize parataxis?
The reason why the which clauses can be considered continuing is that in addition
to meeting the two conditions, they should be finite clauses, which is an impor-
tant precondition. Absolute clauses are non-finite. They are at a lower level in rank
than finite clauses. Therefore, whether absolute clauses can realize parataxis also
requires to be supported by whether non-finite clauses and finite clauses can be
paratactic. In the following, we will try to analyze the functional syntactic struc-
ture of 5-56.
5-56 Frank sat reading the newspaper. (Quirk et al. 1985: 1126)
According to traditional grammar, 5-56 is a simple clause. The predicate sat
cannot introduce an object because it is intransitive, and the present participle
phrase reading the newspaper functions as an attendant circumstance, illustrating
the action state of the event sat. However, according to SFL, there are two differ-
ent processes in 5-56. One is realized by a finite element and the other by a non-
finite element. “So it is a clause complex when embedding is excluded. Here the
relator and the process are conflated and they are realized together by reading”
(Yang 2003: 59). However, Yang (2003) has not discussed the relation between
the two clauses in this clause complex. In Halliday’s thinking, a non-finite clause
can only realize hypotaxis in a clause complex. The non-finite clause in 5-56 real-
izes additive extension. In hypotaxis, two clauses can be translocated, but the two
clauses in 5-56 cannot. This is because if translocated, the non-finite clause read-
ing the newspaper will realize enhancement. The logico-semantic relationship will
not change when a dependent clause is translocated with the dominant clause. This
proves that the clause complex consisting of a finite clause and a non-finite clause
in 5-56 realizes paratactic extension.
Parataxis is different from hypotaxis in that the former “is logically (i) sym-
metrical and (ii) transitive” (Halliday and Matthiessen 2014: 452). When we say
the two clauses in a paratactic clause complex cannot be translocated, we mean the
initiating clause always occurs first. Absolute clauses are always secondary. If an
absolute clause can occur in front of the primary clause, it is certainly dependent.
It should be noted that the distinction of relation types of clause complexes in
SFL is meaning-based and that in traditional grammar is form-based. Seen from
the formal perspective, “subordinate clauses are finite, non-finite, or verbless”
(Greenbaum 1996: 306), but “co-ordination and subordination can sometimes
express similar meaning relationships” (ibid). The relationship between the finite
clause and the non-finite clause in 5-56 is subordination in form, while parataxis in
meaning.
5.5 Continuing Absolute Clauses 97
From the above analysis, it is clearly seen that the independence of absolute
clauses changes with the depth of embedding. There are no absolutely independent
absolute clauses. Finite clauses are in principle independent and have a tendency
to be used alone, while non-finite clauses are dependent in principle. Absolute
clauses are non-finite clauses. The so-called absolute is relative to the primary
clause to which an absolute clause is attached. In the three types of absolute con-
structions distinguished by traditional grammar, the subject of a free adjunct con-
struction and that of the primary clause can be co-referential, hence a related free
adjunct construction, or not be co-referential, hence an unrelated free adjunct con-
struction, or dangling construction. It is obvious that the two types of free adjunct
constructions are not the same in independence.
Augmented absolute constructions are not the absolute constructions in the
strict sense because they are introduced by the conjunctive preposition with.
Absolute clauses are not introduced by conjunctive expressions and their sub-
jects are not co-referential with those of the primary clauses. Therefore, absolute
clauses are more independent compared with the other two types of absolute con-
structions. However, absolute clauses of different syntactic functions are differ-
ent in independence. In this section, we will discuss the independence of absolute
clauses and their tendency to be independent.
98 5 Relationships Realized by Absolute Clauses
5.6.1.1 Dimension of Subject
The independence of the hypotactic clauses in 5-59 increases with the decreasing
requirement of the subject.
5-59a. Since I left my hospital position, I have made my living by writing.
(COCA_MAG)
b. When questioned, Ray evaded and changed the subject. (COCA_FIC)
c. On reaching Bongi Street, the vehicle was hit violently by a light truck.
(COCA_ACAD)
d. Judging from the way he rode the train, the guy knew martial arts.
(COCA_FIC)
Free adjunct
Augmented absolute
The hypotactic clause in 5-59a is finite, and the subject cannot be omitted. The
conjunction when in 5-59b realizes relator, and the logical subject of the hypo-
tactic non-finite clause is the nominative case and is co-referential with that of
the primary clause. The preposition on in 5-59c realizes relator, and the omitted
subject may be the oblique case, as in on him reaching Bongi Street, or posses-
sive, as in on his reaching Bongi Street. Although “the preferred form in current
usage is the ‘oblique’ case” (Halliday and Matthiessen 2004: 421), the requirement
of a subject of this kind of non-finite clauses has clearly decreased. The subject
of a dependent non-finite clause is usually co-referential with that of the primary
clause and hence is typically omitted. The non-finite clause in 5-59d has not a con-
junctive expression, and its subject is not co-referential with that of the primary
clause. The omitted subject may be we or a general pronoun one. Crystal (2008: 2)
refers to this kind of constructions independent from the primary clauses as abso-
lute constructions, but not absolute clauses for lacking an explicit subject.
5.6.1.2 Dimension of Relator
The independence of the hypotactic clauses in 5-60 increases with the decreasing
requirement of the conjunctive expression.
5-60a. Since I’ve come back, I haven’t worked a day. (COCA_NEWS).
b. Because of bylaws requiring a window, he decided to leave an open,
32,000 square foot inner core. (COCA_NEWS)
c. With money being so tight, you have to shop around. (COCA_NEWS)
d. Compromise having failed, there was left only force. (BNC_ACAD)
The finite secondary clause in 5-60a has the strongest requirement of a rela-
tor. If there is not an explicit conjunction, the relation between the two clauses is
not hypotactic, but paratactic. Despite that the non-finite clause in 5-60b has its
own subject, the relator is realized by the preposition because of which makes the
dependent clause not being able to be independent from the dominate clause in
structure. If the preposition is removed, the non-finite clause becomes an absolute
clause. In 5-60c, as in 5-60b, the non-finite clause with subject is introduced by a
preposition with, but with “does nothing to narrow down the range of logical roles
which it may assume” (Stump 1985: 13). It only realizes relator which prevents
the non-finite clauses with subject from forming absolute clauses. The non-finite
clause with subject in 5-6d is not introduced by an explicit conjunctive expression
and is independent from the primary clause in structure, hence forming an abso-
lute clause. The logico-semantic relation can be inferred from the meaning of the
primary clause. For example,
5-61a. John went off by himself, the rest of us staying behind. (Halliday and
Matthiessen 2004: 404)
b. Lu found an empty one and got in, her mother following. (COCA_FIC)
100 5 Relationships Realized by Absolute Clauses
The primary clause in 5-61a emphasizes that John leaves by himself, and the abso-
lute clause functions as the explanation of this proposition. The logico-semantic
relation is elaboration. The absolute clause in 5-61b supplements the information
of the primary clause, and the logico-semantic relation is extension.
Absolute clauses can form clause complexes with finite clauses realizing elab-
oration, extension, and enhancement. The difference between an absolute clause
and a subjectless non-finite clause is that the former is more independent. It may
not be introduced by a conjunctive expression and can be used independently with
all the basic characteristics of parataxis. In Chap. 6, we will carry out a qualitative
and quantitative research of the relationships realized by absolute clauses.
The dependence of dependent clauses increases along two lines starting from
finite clauses. One is the decreasing of the requirement of a subject and the other
is the decreasing of the requirement of a conjunctive expression. The first line
ends at absolute non-finite clauses, i.e., dangling constructions in traditional gram-
mar, and the second line ends at absolute clauses. Both belong to the category of
absolute constructions. The difference between the two is that absolute clauses
are semantically not attached to any element of the primary clause and are not
intended to modify any nominal group. An absolute non-finite clause, however, is
purposed to modify a certain nominal group, but is wrongly linked up with another
nominal group.
In clauses composed of nominal groups, verbal groups and adverbial groups, the
adverbial groups realizing circumstances are relatively loosely connected with
the verbal groups realizing processes, and their positions are relatively flexible. A
clause without circumstances is still structurally integrated. In clause complexes
with absolute clauses realizing circumstances, the absolute clauses have no formal
links with the primary clauses and they are not indispensible in meaning either.
Therefore, absolute clauses of circumstance are strong in independence. They are
usually separated from the primary clauses with a comma in form and are also
flexible in position: An absolute clause can be preceding or following the primary
clause, or even in between the primary clauses. Without the absolute clause, the
primary clause still enacts a complete proposition or proposal.
5.6 Independence of Absolute Clauses 101
One is that the subject of the non-finite clause is not co-referential with that of
considers, so it requires its explicit subject. Even if the subject of the non-finite
clause is co-referential with that of considers, the non-finite clause still requires
an explicit subject, the slot of which can only be filled with the reflexive pronoun
himself. The other is that what the verb considers projects is a fact, but not an act.
However, infinitives are perfective and hence cannot construe a fact. It is only
when they have their own subject that they can construe a complete proposition of
fact. For example,
5-65a. Sidney considers trying hard to be good spirited. (COCA_FIC)
b. *Sidney considers to try hard to be good spirited.
The subject of the non-finite clause and that of the main verb of the project-
ing clause are co-referential in 5-65. The gerund in 5-65a is imperfective and
can construe a fact proposition, functioning as the complement of the main verb
considers. The infinitive in 5-65b is a projected act, and so it cannot construe a
fact proposition, nor function as the complement of the main verb considers. This
further indicates that although a project act can also form an absolute clause, the
independence of a projected fact is stronger.
The relationship between the subject of non-finite clauses and the main verb of
the primary clauses is a vital factor to affect the independence of absolute clauses.
5-66a. He wanted me to be a lawyer. (COCA_FIC)
b. He invited me to stay for a few days. (COCA_MAG)
c. To stay for a few days, he invited me.
The word me in 5-66a is less closely tied with the main verb in the project-
ing clause than that in 5-66b. The former first of all functions as the subject of
the infinitive, and the construction as a whole functions as the complement of the
main verb in the projecting clause and hence is more independent. The latter first
of all functions as the complement of the main verb in the projecting clause and
the infinitive to stay for a few days can be considered as the purpose of the main
verb, realizing enhancement. The whole structure can be changed into 5-66c.
As can be seen from 5-64 and 5-66, projected clauses with subject are possible
to form absolute clauses, the independence of which is determined by the relation
of the subject of the non-finite clause and the main verb. The relation between the
subject of the non-finite clause in 5-66b and the main verb is the weakest, and
therefore, the independence of the non-finite clause with subject is the strongest.
The subject of the non-finite clause in 5-66a belongs only to the non-finite clause,
but this non-finite clause with subject construes an act rather than a fact and the
independence is relatively weaker. The subject of the non-finite clause in 5-64a is
also the complement of the main verb heard, but this complement is not always
necessary. For example, changing 5-64a into The wind rustling in the bushes I
heard is acceptable, while changing it into Rustling in the bushes I heard the wind
is unacceptable. From this point, the non-finite clause in 5-64a is still an abso-
lute clause, except that the independence is much weaker. The non-finite clause
in 5-66b is also a subject + predicate construction, but it cannot form an absolute
5.6 Independence of Absolute Clauses 103
clause because me cannot be separated from the main verb. Thus, the non-finite
clauses with subject in the four example sentences are on a cline in independence:
5-64b > 5-66a > 5-64a > 5-66b
Seen from the process types realized by verbs, non-finite clauses projected by ver-
bal verbs cannot form absolute clauses and those projected by mental verbs can
form absolute clauses. This fits well with the independence rules of the projected
phenomenon: Verbal verbs project acts, mental verbs project facts, and projected
facts are more independent than projected acts.
Absolute clauses functioning as subject and appositive are both fact clauses.
Both are more independent than absolute clauses functioning as complement. As
for subject and appositive, the former is more independent than the latter. The rea-
son is that although the two are both embedded clauses, the subject of the latter
is still controlled by the main verb to some extent. For example, if discipleship
in 5-64b is a personal pronoun, it should be accusative case. The subject of the
former is not controlled by the main verb; it can be nominative case and hence is
more independent.
According to this analysis, we can make this conclusion: Seen from the inde-
pendent tendency, different relation types of absolute clauses form a cline:
Elaboration > Extension > Enhancement > Embedding > Projection (Fact > Act)
In constructions containing two clauses, the two clauses are interdependent. When
one of the two clauses is an absolute clause, its independence differs in terms of
the different relations between clauses. The independence of absolute clauses is
also embodied in that of the primary clauses.
In constructions containing absolute clauses, the absolute clauses can realize
circumstances and participants as well. “The configuration of process + partici-
pants constitutes the experiential centre of the clause.” (Halliday and Matthiessen
2004: 176) Circumstances are peripheral and they are not involved in the pro-
cesses. The interdependence between adverbial groups realizing circumstances
and verbal groups realizing processes is relatively weak. The absence of the cir-
cumstance will not affect the integrity of the clausal structure. That is to say, in
constructions containing absolute clauses realizing circumstances, the primary
clauses are less dependent on the absolute clauses. Similarly, in a clause com-
plex of parataxis, each of the two clauses has its own syntactic structure, so dou-
ble transitivity analyses is not allowed. Since absolute clauses of extension can
be transformed into continuing clauses, in clause complexes containing absolute
clauses of extension, the primary clauses is the least dependent on the absolute
clauses and hence is the most independent. In clause complexes with absolute
clauses of enhancement, the absolute clauses can be transformed into finite
clauses, realizing the circumstance of the main verb, and the whole construction
can be double-analyzed in transitivity. Removing the absolute clauses will affect
104 5 Relationships Realized by Absolute Clauses
the basic configuration of the primary clauses, but the structure is still complete.
Therefore, the primary clauses are still independent of the absolute clauses.
Compared to circumstances, participants are relatively closer to processes in
relation. In the unmarked case, participants are closely bound to the processes
and hence are indispensible. From the perspective of the relationship between the
subject and the main verb, absolute clauses functioning as subject are more inde-
pendent of the primary clauses than those functioning as complement, but from
the requirement of participants by the main verbs, the constructions with abso-
lute clauses functioning as subject are less independent than those with absolute
clauses functioning as complement. Here is the reason. Although the subject and
complement of a verbal group are both indispensible, in the same syntactic struc-
ture, the same constituents can be omitted. This is to say, both the subject and the
complement can be omitted. However, the omission of the subject is structural,
which can only be realized in paratactic clause complexes, while the omission of
the complement may be cohesive. This shows that a verb requires the subject more
intensely than the complement, and hence, a clause without complement is more
independent than that without subject. When the omitted elements are obvious, a
clause without complement can be used alone, but a clause without subject cannot.
For example,
5-67a. He stopped and took off his watch. (COCA_FIC)
b. Millions of people watch the event (Sun., March 21, 8 p.m., ABC) on TV
around the world. Who will win and who deserves to win? (COCA_NEWS)
5-67a is a paratactic clause complex. The subject of the second clause is co-
referential with that of the first clause and hence is omitted. 5-67b consists of two
separate sentences, the second sentence of which is a paratactic clause complex.
The complement the event in both of the two clauses in this clause complex is
omitted. Ellipsis is a cohesive device at the textual level.
Appositive clauses are not direct participants of the main verbs, and so their
relationship with the main verbs is relatively loose. In constructions with absolute
clauses of appositive, the absolute clauses are strongly independent. The clauses
within which they are embedded are also strongly independent. The independence
of different kind of primary clauses is also on a cline:
Elaboration > Extension > Enhancement > Projection > Embedding
Despite the fact that the subject of an absolute subject clause has nothing to do
with the main verb, this kind of absolute clause is still strongly independent.
However, the intense requirement of a subject by the main verb weakens the inde-
pendence of the construction as a whole. It is obvious that of the two types of
independent tendencies, the primary clause-based independent tendency is more in
line with the characteristics of absolute clauses.
5.7 Summary 105
5.7 Summary
References
Biber, D., Johansson, S., Leech, G., Conrad, S., & Finegan, E. (1999). Longman grammar of spo-
ken and written English. London: Longman.
Crystal, D. (2008). A dictionary of linguistics and phonetics (6th ed.). Oxford: Blackwell
Publishing.
Curme, G. O. (1931). A grammar of the English language: Syntax. Boston: D. C. Heath.
Fawcett, R. P. (1996). A systemic functional approach to complementation in English. In
M. Berry, C. Butler, R. P. Fawcett & G. W. Huang (Eds.). Meaning and form: Systemic func-
tional interpretations (pp. 297–366). Norwood, NJ.: Ablex.
Greenbaum, S. (1996). The Oxford English grammar. Oxford: Oxford University Press.
Halliday, M. A. K. (1966[2002]). The concept of rank: A reply. In J. Webster (Ed.), Collected
works of M. A. K. Halliday, vol. 1: On grammar (pp. 118–126). London: Continuum.
Halliday, M. A. K. (1994). An introduction to functional grammar (2nd ed.). London: Edward
Arnold.
Halliday, M. A. K., & Matthiessen, C. M. I. M. (2004). An introduction to functional grammar
(3rd ed.). London: Edward Arnold.
Halliday, M. A. K., & Matthiessen, C. M. I. M. (2014). Halliday’s introduction to functional
grammar (4th edn.). London and New York: Routledge.
He, W. (2002). On Recursiveness, Embedding and Functional Syntactic Reanalysis. Foreign
Language Research, 3, 64–69.
Huang, G.-W. (1998a). Form is the realization of meaning. Foreign Languages and Their
Teaching, 9, 4–7.
Huang, G.-W. (1998b). A functional analysis of the English causative structure. Journal of
Foreign Languages, 1, 12–16.
Huang, G.-W. (1998c). A functional analysis of the “wh—continuing clause” in English. Modern
Foreign Languages, 1, 1–9.
106 5 Relationships Realized by Absolute Clauses
Huang, G.-W. (1999). Explorations in English language and linguistics. Guangzhou: Zhongshan
University Press.
Quirk, R., Greenbaum, S., Leech, G., & Svartvik, J. (1985). A comprehensive grammar of the
English language. London and New York: Longman.
Reuland, E. J. (1983). Governing -ing. Linguistic Inquiry, 14, 101–136.
Stump, G. T. (1985). The semantic variability of absolute constructions. Dordrecht: D. Reidel
Publishing Company.
Thompson, G. (1996). Introducing functional grammar. London: Edward Arnold.
Yang, B.-J. (2003). A study of non-finite clauses in English: A systemic functional approach.
Beijing: Foreign Language Teaching and Research Press.
Zhang, Z.-B. (1997). A new English grammar (3rd ed.). Shanghai: Shanghai Foreign Language
Education Press.
Chapter 6
Absolute Clauses Distributed
in Three Corpora
6.1 Introduction
6-4a. When the emperor saw a flagstone engraved with a cross he ordered it to be
raised (BNC_MISC).
b. When the king commanded deer to be taken for his use in the royal
forests it was usually the warden’s responsibility to see that this was done in
the proper manner (BNC_NA).
Therefore, of the three types of absolute clauses of projection, we only retrieve
those of reported ideas, including proposition and proposal. There might be too
many concordance lines retrieved, if all the verbal groups projecting ideas are con-
sidered, so we just select several most representative verbs that can project both
finite clauses and absolute clauses, such as believe, consider, expect, feel, find,
hear, know, see, want. Then, we write the following regular expression:
RE7.\S+_(PP|N\w+)\s\S+_V\w*\s\S+_PP\s(\S+_TO\s\S+_V[BDHV]|\S
+_V(B|D|H|V)[GN])
Using this regular expression, we have collected 533 absolute clauses of projec-
tion from the Brown Family Corpora. In the next subsections, we will observe the
distributions of these absolute clauses.
6.2.1 Overall Distributions
In this section, we will analyze the overall distributions of the 382 absolute clauses
of expansion and the 533 absolute clauses of projection extracted from the Brown
Family Corpora, including the formal distribution, the historical distribution, and
the stylistic distribution.
By means of the regular expressions, we can extract all absolute clauses of expan-
sion with noun or pronoun subject and present participle, past participle, infinitive,
adjective, adverb, preposition phrase or noun predicate. The frequencies of all the
forms of absolute clauses are shown in Table 6.1.
Table 6.1 shows that, out of the 382 absolute clauses of expansion, there are
224 with non-finite verb predicate and 158 with verbless predicate, respectively.
The present participle stands out with 152 instances, accounting for 39.8 % of
the total, with the past participle, preposition phrase, adjective, and adverb fol-
lowing in sequence. The statistics also shows that absolute clauses with infini-
tive and noun predicate are the smallest in number. All the forms except infinitive
can be considered as absolute clauses with the present participle being omitted.
Therefore, some grammarians also refer to absolute clauses in the traditional sense
as absolute participles (see Zandvoort 1972).
The absolute clauses of projection can be classified in form by the projection
verbs selected (see Table 6.2).
110 6 Absolute Clauses Distributed in Three Corpora
6.2.1.2 Historical Distribution
The historical and regional distributions of the 382 absolute clauses of expansion
and the 533 absolute clauses of projection are shown in Table 6.3.
Table 6.3 shows that absolute clauses of expansion are evenly distributed in the
six corpora. Seen from the regional distribution, British English is slightly out-
numbered by American English in 1961 and 1991/1992, while in 2009, they are
6.2 Research Based on Brown Family Corpora 111
equal in number. Seen from the historical perspective, from 1961 to 1991/1992,
both American English and British English are somewhat on the decline, but from
1991/1992 to 2009, both go steadily up with American English ending up increas-
ing to the level of 1961 and British English exceeding the level of 1961. Although
the total number is increasing in the span of nearly 50 years, it does not show any
significant regular tendency in historical and regional distributions.
Absolute clauses of projection are also evenly distributed in both region and
time. British English is slightly outnumbered by American English in 1961
and 1991/1992, while in 2009, American English dominates British English.
Historically, from 1961 to 1991/1992, American English is somewhat on the
decline, but from 1991/1992 to 2009, it goes up over the level of 1961, and British
English is right opposite. In general, neither absolute clauses of expansion nor
those of projection show significant difference in either regional or historical
distributions.
6.2.1.3 Stylistic Distribution
Table 6.4 aims to interpret the distribution of the absolute clauses in such four reg-
isters as press, general prose, learned and fiction.
Table 6.4 shows that the stylistic distribution of absolute clauses varies
greatly, mostly appearing in fiction, with general prose, press and learned fol-
lowing. However, there is little difference between American English and British
English in all the four styles. For example, except that British English outnum-
bers American English by one instance of both expansion and projection and seven
instances of projection in press, in the other three styles in expansion, American
English slightly outnumbers British English, and in the other two styles in projec-
tion, British English outnumbers American English. In general, the stylistic dis-
tribution of absolute clauses shows no significant difference between American
English and British English or between expansion and projection.
In fact, the original frequencies of absolute clauses in the four registers can
hardly bring an adequate explanation that absolute clauses are scarcely used in
press or learned, because in the six corpora, there are relatively fewer words in the
press and learned sub-corpora and as such, the number of the absolute clauses in
250
200
150
100
50
0
Press General Prose Learned Fiction
6.2.2 Functional Distribution
In this section, we will analyze the functional distribution of the 915 absolute
clauses extracted from the Brown Family Corpora, including the regional distribu-
tion, historical distribution and stylistic distribution, etc.
In this section, we will analyze the historical distributions of the function types of
absolute clauses in the Brown Family Corpora (see Table 6.6 and Fig. 6.2).
Table 6.6 and Fig. 6.2 show that absolute clauses of enhancement, elaboration,
and projection nearly keep steady in number, and only those of extension show
a trend to increase. The number of elaboration in 1991/1992 is slightly larger
than that in 1961 but undergoes a decrease in 2009. The number of extension in
1991/1992 is a bit smaller than that in 1961 but undergoes a substantial increase
in 2009. Enhancement is on a gradual decline throughout the nearly 50 years.
Although the overall frequency of absolute clauses manifests no distinct changes
along time, yet from the perspective of function types, the total number of the
absolute clauses of extension is on the increase and that of enhancement is on the
decrease.
Figures 6.3 and 6.4 show that the four functions of absolute clauses share the
similar distribution trend along time in British English and American English. The
number of absolute clauses of enhancement in American English is on a steady
decrease, and that of extension in British English keeps increasing but that in
114 6 Absolute Clauses Distributed in Three Corpora
200
180
160
140
120
100
80
60
40
20
0
1961 1991/1992 2009
Enhancement 21 18 14
Extension 73 63 101
Elaboration 33 36 23
Projection 183 164 186
American English goes at first downward and then upward. An irregular distri-
bution trend is shown in projection. In British English, it goes downward, and in
American English, first downward and then upward.
In this section, we will analyze the stylistic distribution of the four function types
of absolute clauses.
For a good visual presentation, we change the total numbers of enhancement,
extension, and elaboration into 533, the total number of projection (see Fig. 6.5).
Table 6.7 and Fig. 6.5 show that these four function types of absolute clauses
differ greatly in stylistic distribution. All function types of absolute clauses tend
6.2 Research Based on Brown Family Corpora 115
100
90
80
70
60
50
40
30
20
10
0
1961 1991/1992 2009
Enhancement 12 11 11
Extension 29 32 54
Elaboration 17 12 4
Projection 87 89 78
120
100
80
60
40
20
0
1961 1991/1992 2009
Enhancement 9 7 3
Extension 44 31 47
Elaboration 16 24 19
Projection 96 75 108
to be used in fiction, with extension dominating the other three and projection
following, and none of the four functions tends to occur in learned, especially
extension and projection. This is right opposite in fiction. Of the four functions,
the number of extension is the smallest in press, general prose, and learned. In
press and learned, enhancement counts the most and in general prose, elaboration.
The number of elaboration is the largest in general prose and the second largest in
learned. This is because fiction is characterized with narration of events and depic-
tion of actions and hence advisable to use absolute clauses of extension, while
learned is characterized with argumentation and hence favorable to use absolute
116 6 Absolute Clauses Distributed in Three Corpora
450
400
350
300
250
200
150
100
50
0
Press General Prose Learned Fiction
Enhancement 91 151 70 221
Extension 31 72 9 421
Elaboration 35 197 52 249
Projection 55 129 10 339
Traditional grammar fails to reach a final agreement on the case of the personal
pronoun subject of absolute clauses. It is popularly accepted that the subject of an
absolute clause is a zero noun or a nominative pronoun, or occasionally an accu-
sative pronoun. This at least shows that it is uncommon for accusative pronouns
to function as the subject of absolute clauses. In this section, we will analyze
the distribution of the personal pronoun subject of absolute clauses and its case
choice, including the historical distribution and the stylistic distribution. The case
of the personal pronoun subjects of absolute clauses of projection is the assigned
6.2 Research Based on Brown Family Corpora 117
accusative by the projection verbs, so we count only the personal pronoun subjects
of absolute clauses of expansion which are not assigned case by the main verbs in
the matrix clauses.
Historical Distribution
Among the 382 absolute clauses of expansion extracted from the Brown Family
Corpora, there are merely 13 instances with personal pronoun subject. This is not
enough to conduct statistical analyses. Therefore, it is necessary to appropriately
relax the restrictions on the retrieval and to rewrite the following regular expres-
sion by which all the collocations consisting of a personal pronoun and a non-
finite or verbless element can be retrieved.
RE8.\S + _(,|SENT|:)\s\S + _PP(\s\S + _RB\w*)*\s(\S + _TO\s\S + _V[BDHV]|\S
+ _(V(B|D|H|V)[GN]|JJ\w*|IN|DT|N\w +))
Using this regular expression, we extracted 61 absolute clauses with personal
pronoun subject from the Brown Family Corpora, including 31 in American
English and 30 in British English. Table 6.8 shows the historical and regional dis-
tributions of the 61 absolute clauses. The number of finite clauses with personal
pronoun subject extracted from the Brown Family Corpora reaches 143,568. The
ratio of absolute clauses with personal pronoun subject and the finite clauses is
4.25 to 10,000. This indicates that personal pronouns are not commonly used as
the subject of English absolute clauses. This is mainly because personal pronouns,
as anaphora in the precedent clauses, usually refer to a nominal group which is
usually a participant of the main verbs in the primary clauses. In the four types of
probable subjects distinguished by Kortmann (1991: 92), the co-referential sub-
jects of absolute clauses and the primary clauses are mostly personal pronouns,
and absolute clauses with personal pronoun subject mostly realize elaboration, that
is, absolute appositive clauses.
Table 6.8 shows that the overall distribution of absolute clauses with personal
pronoun subject is relatively uniform. There are only two more instances in
1991/1992 than in 1961, and in 2009, the total number remains the same. In addi-
tion, the number in American English is larger than that in British English in both
1961 and 1991/1992, and even is on a slightly increase in 1991/1992, but in 2009,
it undergoes a clear decrease, which is outnumbered by British English. Although
the number in British English remains the same in 1961 and in 1991, it is smaller
than that in American English, and in 2009, it increases and outnumbers American
English. Nevertheless, on the whole, the distribution manifests no distinct regional
or historical changes.
The distribution of the 61 personal pronoun subjects in the four registers of press,
general prose, learned, and fiction is shown in Table 6.9.
Table 6.9 shows significant register differences of absolute clauses with per-
sonal pronoun subject. Such absolute clauses are mostly used in fiction and gen-
eral prose. In fiction, American English outnumbers British English, but in press
and general prose, British English takes the lead in number. In learned, abso-
lute clauses with personal pronoun subject are inactive, with only two inanimate
third person reflexive pronouns in American English and none in British English.
Undoubtedly, these original data cannot fully explain the fact that absolute clauses
with personal pronoun subject are ill-performed in press and learned, because the
vocabulary of the two registers in the Brown Family Corpora is smaller, and so
the absolute number of personal pronouns is also smaller. Therefore, to facilitate
effective comparison, the original frequency will also be converted into the stand-
ard frequency of per million words, as is shown in Fig. 6.6.
30
20
10
0
Learned Press General prose Fiction Average
Fig. 6.6 Stylistic distribution of personal pronoun subjects of absolute clauses (per million words)
6.2 Research Based on Brown Family Corpora 119
Figure 6.6 shows that American English shares the same frequency with British
English in terms of personal pronoun subjects of absolute clauses. The frequency
reaches the highest in fiction which outnumbers the total of the other three styles.
In fiction and learned, American English exceeds British English in number, while
in press and general prose, American English falls behind. In American English,
only fiction stands over the average level and gains the dominant advantage
against the other three styles which are evenly distributed nearly to form a straight
horizontal line shown in the figure. Meanwhile, in British English, the four styles
of learned, press, general prose, and fiction present an arithmetic increase trend in
sequence, with general prose and fiction standing over the average level, which
implies that in comparison with American English, British English tends to use
absolute clauses with personal pronoun subject in particular in general prose.
Regional Distribution
Among the 61 instances of the absolute clauses with personal pronoun subject,
except 12 unmarked cases such as absolutive you and it as well as their corre-
sponding reflexives, there are 49 instances of case-marked first person and third
person pronouns, including 28 nominatives, 12 accusatives, and 9 reflexives, as is
shown in Table 6.10.
Table 6.10 shows that the case of personal pronoun subjects of absolute clauses
shares similar trend in both American English and British English. Nominatives
count the most, then accusatives follow, and reflexives stand the least. In addition,
the case mark of reflexives is usually accusative, because in finite clauses, reflex-
ives can alone be used as object rather than subject. Therefore, of the 49 case-
marked personal pronouns, there are actually 28 nominatives and 21 accusatives,
as is shown in Table 6.11.
Historical Distribution
Stylistic Distribution
Table 6.14 Stylistic distribution of the case of personal pronouns in the Brown Family Corpora
Press General prose Learned Fiction Total
Nominative 14,238 33,070 5,094 62,325 114,727
Accusative 5,596 16,891 2,376 33,896 58,759
Total 19,834 49,961 7,470 96,221 173,486
500
400
300
200
100
0
Learned Press General prose Fiction Average
As can be seen, among these four styles, there are altogether 173,486 case-
marked first and third person pronouns, including 114,727 nominatives and 58,759
accusatives, the ratio being 1:0.512. In addition, in these four styles, nomina-
tives take an overall advantage over accusatives. Of the 49 case-marked personal
pronouns, nominatives and accusatives are of the ratio 1:0.75. According to the
overall occurrences of the case of personal pronouns, we can figure out the stylis-
tic distribution of the case of personal pronoun subjects of absolute clauses (see
Fig. 6.7).
Table 6.13 shows that the 49 case-marked personal pronouns are mainly in fic-
tion and general prose, and fiction exceeds general prose in number. This is mainly
because there are more occurrences of personal pronouns in fiction and general
prose than in the other two registers. However, as is shown in Fig. 6.7, after con-
verted into the standard frequency of per million personal pronouns, the scene has
greatly changed. In other words, general prose conversely stands out from fiction,
and even to such an extent that the average frequency of fiction is slightly lower
than the total average frequency. Figure 6.7 also shows that in different styles,
the distributions of nominatives and accusatives differ greatly. In the four styles,
accusatives take the significant advantage against nominatives. In general prose,
there is the widest gap between nominatives and accusatives, while in fiction,
122 6 Absolute Clauses Distributed in Three Corpora
the proportion of nominatives to accusatives keeps steady with the total average
frequency. It is only in press that accusatives fall behind nominatives, and even in
press, there are more nominatives than in fiction. It should be noted that personal
pronouns as well as absolute clauses with personal pronoun subject are seldom
used in learned. In the six learned sub-corpora, there are only two absolute clauses
with itself as subject.
The above statistics indicates that when personal pronouns have to be used as
the subject of absolute clauses, accusative personal pronouns are more likely to be
taken rather than nominative personal pronouns.
6.2.4 Built-in Problems
From the above quantitative research, we can draw the following conclusion. The
distribution of absolute clauses manifests no distinct regional or historical differ-
ences in the Brown Family Corpora, but manifests significant stylistic differences.
Among the three types of expansion, the number of absolute clauses of extension
is the largest, while that of enhancement, the smallest. On the whole, the total
number of the absolute clauses of enhancement goes on a steady decline, while
that of extension keeps increasing. In terms of the stylistic distribution, absolute
clauses mainly occur in fiction, but seldom in learned. This is obviously not in
agreement with the traditional view that absolute clauses are mostly used in formal
texts, because from the perspective of formality, learned is the most formal, then
press, general prose and fiction follow in succession. However, among these four
styles, the distribution of absolute clauses is right the other way, with fiction the
most, and then come general prose, press, and learned in turn.
In order to demonstrate the feasibility of the analysis on stylistic distribution, a
comparison will be drawn between the overall frequency of absolute clauses and
the frequency of absolute clauses with personal pronoun subject. To facilitate the
effective comparison, we convert the number of absolute clauses with personal
pronoun subject from 61 to 382 (see Fig. 6.8).
Figure 6.8 shows that absolute clauses share almost the same stylistic distribution
trend with absolute clauses with personal pronoun subject. From learned to press
to general prose, both are on steady increase, but the increasing speed of absolute
clauses with personal pronoun subject is a bit higher. However, from general prose
to fiction, both increase dramatically, and the overall frequency of absolute clauses
increases faster than that of absolute clauses with personal pronoun subject. This
indicates that absolute clauses tend to be applied in fiction rather than in learned. In
addition, there are no spoken data in the Brown Family Corpora, and among these
four registers, fiction is the closest to spoken. However, among the four registers,
fiction is the richest in absolute clauses. Therefore, we have the first question to
consider.
6.2 Research Based on Brown Family Corpora 123
180
160
140
120
100
80
60
40
20
0
Learned Press General prose Fiction
Overall Frequency 20.08 27.47 32.77 166.67
Personal Pronoun 13.09 29.69 55.73 132.26
Q1: Can this distribution indicate that absolute clauses tend to be used in
informal texts?
The above quantitative analysis shows that neither in terms of the overall fre-
quency nor in terms of the frequency of the personal pronoun subject, the distri-
bution of absolute clauses shows distinct regional or historical differences. With
respect to historical distribution, the span of time of the Brown Family Corpora is
no longer than 50 years, within which the evolution tendency of language is not
always sensitive. The historical distribution research of absolute clauses might not
have caught up with its own internal trend of historical variation. Therefore, we
have the second question here.
Q2: Can absolute clauses manifest historical differences in a longer span of time?
In order to answer the questions, we will observe language facts from the BNC
and COCA in the following sections.
6.3 BNC-Based Observations
In this section, we will use the BNC to carry out a further quantitative research
on the stylistic distribution of absolute clauses. The research is mainly about the
application of absolute clauses in both informal and formal English.
124 6 Absolute Clauses Distributed in Three Corpora
6.3.1 Overall Frequency
20000
18000
16000
14000
12000
10000
8000
6000
4000
2000
0
Spoken Newspaper Magazine Non-ac Misc Academic Fiction
Expansion 361 430 1776 952 1238 489 7763
Projection 11391 3927 4035 2680 4012 1944 18429
35000
30000
25000
20000
15000
10000
5000
0
Spoken Newspaper General prose Academic Fiction
Brown Expansion 3089 3700 2360 18674
Brown Projection 4284 4309 861 18368
BNC Expansion 1381 1643 4661 1869 29661
BNC Projection 11391 3927 3523 1944 18429
clauses to finite clauses. It is also noticed that in the Brown Family Corpora, there
are much fewer absolute clauses of projection in academic than those of expansion
and those of expansion and projection in the BNC. This is because we collected only
the absolute clauses with personal pronoun subject and pronouns are not preferred
in academic. This is also true for absolute clauses of projection in the BNC although
the gap is not that sharp. The four distribution lines in Fig. 6.10 are nearly over-
lapped, but some obvious differences can still be noticed. For instance, the number
of expansion in fiction in the BNC is larger than that in the Brown Family Corpora.
This comparison can at least explain that the frequency of absolute clauses in
the Brown Family Corpora and that of projection in the BNC are the lowest in for-
mal academic, and it is only slightly higher than that in spoken and that in news-
paper in the BNC. According to the data in the BNC, although absolute clauses do
not occur considerably in the formal academic, they are really rarely seen in the
informal spoken texts.
6.3.2 Functions
Although absolute clauses of elaboration are not included in the traditional sense
of absolute clauses, this kind of construction can also be considered as shrunken
finite clauses as other two function types of absolute clauses. To guarantee the
validity of the data analyses, we take absolute clauses of extension and enhance-
ment and those of projection in the BNC as examples to analyze the tendency to
choose absolute clauses from attendant circumstances introduced by with, from
6.3 BNC-Based Observations 127
Table 6.19 Stylistic distributions of with and absolute clauses of extension (per hundred million
words)
Spoken Newspaper General prose Academic Fiction Total
With construction 23,706 74,744 87,757 68,362 76,402 330,971
Absolute clauses 271 344 823 235 6,499 8,172
300000
250000
200000
150000
100000
50000
0
Spoken Newspaper General prose Academic Fiction
with construction 23706 74744 87757 68362 76402
Absolute clauses 10976 13932 33332 9518 263213
400
350
300
250
200
150
100
50
0
Spoken Newspaper General prose Academic Fiction
Ratio 46.3 18.64 38 14 345
Fig. 6.12 The ratio of the absolute clauses of extension to with constructions (equal totality)
Figure 6.15 shows that when the total number of absolute clauses of projection
is converted to that of complement clauses introduced by the complementizer that,
absolute clauses outnumber that clauses in spoken, general prose, and fiction, with
the gap in fiction being the largest and that clauses outnumber absolute clauses in
newspaper and academic. Figure 6.16 shows the ratio of absolute clauses of pro-
jection to that clauses in each of the five styles.
In Fig. 6.16, we can see that absolute clauses of projection are rather prefer-
able in spoken, general prose, and fiction, while that clauses are preferable in
130 6 Absolute Clauses Distributed in Three Corpora
600000
500000
400000
300000
200000
100000
0
Spoken Newspaper General prose Academic Fiction
Subordinators 242923 176737 268246 283537 267020
Absolute clauses 71864 182056 316203 110192 558147
250
200
150
100
50
0
Spoken Newspaper General prose Academic Fiction
Ratio 29.58 103.01 117.878 38.86 209.03
newspaper and academic. This suggests that absolute clauses of neither expansion
nor projection tend to occur in formal academic.
In terms of the two function types of absolute clauses distinguished by tradi-
tional grammar, corpus statistics shows that the number of absolute clauses of
enhancement is significantly smaller than that of extension. Since there are no
explicit conjunctions, the logico-semantic relationship of absolute clauses with
the primary clauses is sometimes unclear, and hence, there may be different
6.3 BNC-Based Observations 131
80000
70000
60000
50000
40000
30000
20000
10000
0
Spoken Newspaper General prose Academic Fiction
that clauses 42414 48183 11688 17187 31918
Absolute clauses 43976 15161 13601 7505 71147
250
200
150
100
50
0
Spoken Newspaper General prose Academic Fiction
Ratio 103.68 31.47 116.37 43.67 222.9
interpretations. Putting the two function types of absolute clauses together, we will
discuss their overall stylistic distributions.
We refer to both with and subordinators as linkers. Then, the ratio of absolute
clauses to linkers in each style will be figured out (see Table 6.22; Fig. 6.17).
We can see from Table 6.22 and Fig. 6.17 that, with equal frequencies, absolute
clauses outnumber linkers only in fiction. In the other four styles (spoken, newspaper,
general prose, and academic), absolute clauses are fewer than linkers. These results
show that absolute clauses differ significantly in stylistic distribution. According to
132 6 Absolute Clauses Distributed in Three Corpora
450
400
350
300
250
200
150
100
50
0
Spoken Newspaper General prose Academic Fiction
Ratio 22.4 33.17 53.24 15.81 388.91
the absolute number and the ratio to linkers, absolute clauses tend to occur in fiction
the most, but they are not preferable in academic. Since there are fewer hypotactic
clauses of extension and enhancement in spoken, absolute clauses are small in num-
ber in spoken. Non-finite projected proposal is preferable in spoken than in newspa-
per and academic, so there are more absolute clauses of projection in spoken than in
newspaper and academic. This will be further discussed in Chap. 7.
6.4 COHA-Based Observations
In terms of the overall frequency of expansion in COHA, there are much more abso-
lute clauses of extension than those of enhancement and elaboration. Such a functional
distribution matches up in all the three corpora, i.e., the Brown Family Corpora,
the BNC and COHA. The number of absolute clauses of enhancement in each of the
three corpora is the smallest of the three types of expansion (see Table 6.24).
To facilitate effective comparison, we convert the total numbers of absolute
clauses of expansion retrieved from the Brown Family Corpora and the BNC into
that (8,577) from COHA. We do not include projection because we use different
regular expressions to collect absolute clauses of expansion from the Brown
Family Corpora and from the BNC and COHA, while we use the same regular
expressions to collect absolute clauses of projection. As for absolute clauses of
projection, we convert the total numbers from the three corpora into the standard
frequencies of per hundred million words (see Figs. 6.18 and 6.19).
Figure 6.18 shows that under the condition of equal totality, the number of abso-
lute clauses of enhancement is the smallest in all the three corpora, and of the three
corpora, the number of enhancement is the smallest in the BNC, with the Brown
Family Corpora and COHA following. Absolute clauses of extension are entirely
different, with the BNC the most, then COHA, and the Brown Family Corpora the
134 6 Absolute Clauses Distributed in Three Corpora
7000
6000
5000
4000
3000
2000
1000
0
Brown BNC COHA
Enhancement 1190 505 1213
Extension 5321 6649 5708
Elaboration 2066 1423 1656
12000
10000
8000
6000
4000
2000
0
Brown BNC COHA
Projection 8670 6593 10113
least. Figure 6.19 shows that absolute clauses of projection appear mostly in COHA,
and the number is the smallest in the BNC. This distribution is similar to that of
enhancement, but contradicts that shown in Fig. 6.2, according to which absolute
clauses of projection show a trend of increase along time. The historical distribution
of absolute clauses of projection will be further discussed in Sects. 6.4.2 and 6.4.3.
Generally, the four function types of absolute clauses share similar distribution
trend. In Fig. 6.18, the three lines are nearly parallel, with enhancement in all the
three corpora the lowest and extension the highest. Therefore, absolute clauses
manifest no obvious functional differences in all the three corpora.
It is worth noting that the research on the historical distribution based on the
Brown Family Corpora shows that over a span of time of less than 50 years, the
number of absolute clauses of extension is on the increase, while that of enhance-
ment on the decrease. Comparatively, among these three corpora, the data in
the BNC cover the shortest distance from nowadays, while those in COHA the
greatest distance. Coincidentally, in COHA absolute clauses of enhancement and
projection are the most in number and those of extension the fewest among the
three corpora. However, in the BNC, the scene is right opposite. Absolute clauses
of enhancement and projection are the fewest and those of extension, the most.
In the following, a research on historical distribution of absolute clauses retrieved
from COHA will be conducted to test whether absolute clauses of enhancement and
projection are decreasing in number and those of extension increasing along time.
Figure 6.20 shows the historical distribution of the 8,577 absolute clauses of
expansion and the 41,009 absolute clauses of projection given in Table 6.23 over
20 phases totaling a span of 200 years in COHA. To facilitate comparison, we con-
vert the number of expansion into that of projection.
Figure 6.20 shows that the absolute number of absolute clauses of both expan-
sion and projection in COHA is increasing, with only slight declines on several
phases. However, such an overall distribution can hardly manifest the real evolu-
tion of absolute clauses, for the total vocabulary of COHA is around 406 million
and it keeps changing over phases. The frequency over the phases is converted to
the standard frequency of per hundred million words (see Fig. 6.21) so as to facili-
tate comparison.
Figure 6.21 shows that when the original frequencies are converted into the
standard frequency of per hundred million words, the total number of absolute
clauses has increased. The historical distribution of absolute clauses of projection
is nearly uniformly increasing, while that of expansion is generally increasing with
some fluctuations in several phases. The distribution in the phases from 1810s to
1910s is irregular: The first four phases change alternately, while the last seven
phases are on a steady decline. The phases from 1920s to 2000s are on a grad-
ual increase, during which the three phases from 1920s to 1950s keep flat and the
136 6 Absolute Clauses Distributed in Three Corpora
6000
5000
4000
3000
2000
1000
0
1810s 1820s 1830s 1840s 1850s 1860s 1870s 1880s 1890s 1900s 1910s 1920s 1930s 1940s 1950s 1960s 1970s 1980s 1990s 2000s
Expansion 57 373 1047 1052 1688 1563 1477 1396 1549 1568 1353 2314 2238 2230 2333 2960 3242 3060 4399 5039
Projection 36 388 728 911 1128 1264 1605 1780 1847 2397 2707 2897 2626 2549 2747 2849 2750 2907 3538 3355
18000
16000
14000
12000
10000
8000
6000
4000
2000
0
1810s 1820s 1830s 1840s 1850s 1860s 1870s 1880s 1890s 1900s 1910s 1920s 1930s 1940s 1950s 1960s 1970s 1980s 1990s 2000s
Expansion 4826 5385 7601 6555 10248 9165 7957 6871 7519 7096 5960 9020 9097 9159 9505 12345 13613 12087 15744 17042
Projection 3048 5601 5285 5677 6848 7411 8647 8762 8966 10847 11925 11293 10674 10469 11192 11882 11547 11489 12662 11347
phrase of 1980s decreases. Such a distribution can at least suggest that, absolute
clauses manifest a trend of increase rather than decrease in the evolution of nearly
two hundred years.
6.4.3 Function Types
Then how about the historical distributions of the functions of absolute clauses?
Since the total number of words in the phase of 1810s is the smallest among the 20
phases, the absolute number of absolute clauses is also the smallest accordingly,
and the function distribution of absolute clauses manifests no distinct regularity.
Absolute clauses do show a regular historical distribution from the phase of 1820s.
Therefore, the point of view in the traditional sense that absolute clauses are on a
gradual decrease is only applicable to those absolute clauses of enhancement. As
we can see, the number of absolute clauses of extension is obviously increasing
6.4 COHA-Based Observations 137
16000
14000
12000
10000
8000
6000
4000
2000
0
1810s 1820s 1830s 1840s 1850s 1860s 1870s 1880s 1890s 1900s 1910s 1920s 1930s 1940s 1950s 1960s 1970s 1980s 1990s 2000s
Enhancement 810 3106 3645 2413 4006 2691 2885 1647 2507 1731 1327 1342 952 766 468 558 522 397 513 388
Extension 810 1242 2187 2920 3541 4261 3812 3624 4224 4263 3665 6076 6821 6598 7052 8714 11183 9084 11893 13794
Elaboration 3238 1035 1770 1222 2700 2215 1262 1600 789 1103 969 1603 1322 2082 1987 3071 1907 2606 3337 2862
Projection 3048 5601 5285 5677 6848 7411 8647 8762 8966 10847 11925 11293 10674 10469 11192 11882 11547 11489 12662 11347
Among the 8,577 absolute clauses of expansion, there are 936 with pronoun
subjects, of which 517 common pronouns, 419 personal pronouns. And of the
personal pronouns, there are 289 with case markers, including 228 nominatives
138 6 Absolute Clauses Distributed in Three Corpora
1.2 Nominative
Accusative
1
0.8
0.6
0.4
0.2
0
1820s 1830s 1850s 1860s 1870s 1880s 1890s 1900s 1910s 1920s 1930s 1940s 1950s 1960s 1970s 1980s 1990s 2000s
Fig. 6.23 Proportions of the case distribution of absolute clauses with personal pronoun subject
in COHA
accusatives collected in the phases of 1820s, 1860s, and 1900s. However, after
the phase of 1900s, nominatives start to decrease and gradually lose the dominant
position, while accusatives begin to increase and outnumber nominatives after
the phase of 1990s and take over the dominant position. Although such a result is
not manifested in the Brown Family Corpora-based research, the Brown Family
Corpora-based research does not manifest the dominance of nominative pronouns.
Instead, nominatives and accusatives are evenly distributed along time, with nomi-
native pronouns on a decreasing trend.
Researches based on both the Brown Family Corpora and COHA show that
absolute clauses with accusative pronoun subject are not wrong at all; and they
are on a significant increase in comparison with absolute clauses with nominative
pronoun subject.
7000
6000
5000
4000
3000
2000
1000
0
Enhancement Extension Elaboration
Nominative 1279 3724 3574
Accusative 703 6468 1406
Overall 1213 5708 1656
nominatives and accusatives into the standard frequency of 8,577 and analyze the
functional distributions of personal pronoun subjects, as is shown in Fig. 6.24.
Figure 6.24 shows that in comparison with the overall frequency, nominatives
are preferable in absolute clauses of elaboration and enhancement, while accusa-
tives are preferable in absolute clauses of extension.
The above research on the overall historical distribution of absolute clauses with
case-marked personal pronoun subject shows that the accusatives are on a general
trend of increase, and nominatives, decrease. In this section, we will discuss the
functional distributions of the case-marked personal pronoun subjects of absolute
clauses over time (see Table 6.27; Figs. 6.25, 6.26, and 6.27).
Table 6.27 and Figs. 6.25, 6.26, and 6.27 show that the historical distribution of
the case of absolute clauses has no significant differences in the three functions.
Although accusative pronouns count few among all the three functions, they tend
to replace nominative pronouns.
Nominative Accusative
0.8
0.6
0.4
0.2
0
1810s 1820s 1830s 1840s 1850s 1860s 1870s 1880s 1890s 1900s 1910s 1920s 1930s 1940s 1950s 1960s 1970s 1980s 1990s 2000s
Nominative Accusative
1
0.8
0.6
0.4
0.2
0
1810s 1820s 1830s 1840s 1850s 1860s 1870s 1880s 1890s 1900s 1910s 1920s 1930s 1940s 1950s 1960s 1970s 1980s 1990s 2000s
Nominative Accusative
1
0.8
0.6
0.4
0.2
0
1810s 1820s 1830s 1840s 1850s 1860s 1870s 1880s 1890s 1900s 1910s 1920s 1930s 1940s 1950s 1960s 1970s 1980s 1990s 2000s
The COHA-based research shows that the total number of absolute clauses is
not decreasing but increasing along time. The opinion held by traditional gram-
marians that absolute clauses are on a gradual decrease is true only for those abso-
lute clauses of enhancement, but the number of absolute clauses of extension is
obviously increasing. Even those of enhancement are not disappearing, for the dis-
tribution of absolute clauses of enhancement tends to stay at the same level in the
last 50 years.
142 6 Absolute Clauses Distributed in Three Corpora
6.5 Summary
The Brown Family Corpora-based research does not show significant difference
of regional distribution of absolute clauses. The COHA-based research shows that
over the span of 200 years, the total number of absolute clauses tends to increase
but not decrease. This is mainly manifested in the obvious increase of absolute
clauses of extension. Meanwhile, although the number of absolute clauses of
enhancement is decreasing along time, it is by no means disappearing. The data
of the last 50 years show that the distribution of absolute clauses of enhancement
tends to be leveling off. The reason may be the tendency of grammaticalization of
absolute clauses: being fixed into stereotyped expressions.
Both the Brown Family Corpora-based research and the BNC-based research
show that the stylistic distributions of absolute clauses manifest significant differ-
ences. Among the four styles, i.e., press, general prose, learned, and fiction, in the
Brown Family Corpora, absolute clauses tend most to occur in fiction, and fiction
is the most informal and learned the most formal. However, the fact is that the fre-
quency of absolute clauses is the lowest in learned. This is not in agreement with
traditional grammar. Although the fact that absolute clauses are seldom used in the
typical informal style of spoken complies with traditional grammar, there are few
adverbial clauses introduced by conjunctions in spoken. That is to say, the reason
why absolute clauses are seldom used in spoken is that there are the fewest subor-
dinate clauses in spoken than in other styles.
Although the subject of absolute clauses may be a noun or a pronoun, absolute
clauses with pronoun subject are few in number, and the pronoun subjects show
significant differences in stylistic distribution. The Brown Family Corpora-based
research shows that personal pronoun subjects of absolute clauses are mainly
concentrated in general prose and fiction. However, personal pronoun subjects
of absolute clauses per million personal pronouns gain the highest frequency in
general prose, then in fiction and press in succession. In addition, from the per-
spective of case distribution, accusatives tend to occur in general prose and fiction,
and nominatives in press. The COHA-based research on the historical distribution
of case provides further evidence that the number of accusative personal pronoun
subjects is increasing along time.
References
Kortmann, B. (1991). Free adjuncts and absolutes in English: Problems of control and interpre-
tation. London: Routledge.
Stump, G. T. (1985). The semantic variability of absolute constructions. Dordrecht: D. Reidel
Publishing Company.
Zandvoort, R. W. (1972). A handbook of English grammar (6th ed.). London: Longman.
Chapter 7
Discussions
Form, function, case, stylistic characteristics, and historical evolution, etc., have
been studied by many linguists as far as absolute clauses are concerned. However,
there still exist some significant problems. In Chap. 6, we have conducted a
corpus-based quantitative research on the three function types of absolute clauses,
focusing on the function, case, style, and time aspects. In this chapter, a discussion
will be conducted on the research results.
7.1 Function
and form some logico-semantic relations with these sentences. In another sense,
these constructions have no finite verbs, nor are elliptical clauses formed out of the
consistency of syntactic structures, hence are non-finite clauses. In the mean time,
these constructions have their own subjects, hence are absolute clauses. In 7-1a,
the three absolute clauses function as appositives of the precedent clause I’d done
everything already, realizing elaboration. In 7-1b, the absolute clause is accompa-
nying the event of the precedent clause, realizing extension. In 7-1c, the absolute
clause is the reason of the following clause, realizing enhancement. The logico-
semantic relations realized by these absolute clauses are cohesive rather than
structural. However, these kind of absolute clauses are not included in the absolute
clauses in the traditional sense.
Using the regular expressions we wrote, we extracted 78, 281, and 997 absolute
clauses of expansion used alone from the Brown Family Corpora, the BNC, and
COHA, respectively (see Table 7.1).
Table 7.1 shows that in the three corpora, absolute clauses used alone account
for 11.07 % of the total. Although such absolute clauses do not form clause com-
plexes, they realize such logico-semantic relations of extension and enhancement
of time, reason, condition, concession, and so on, and they can also realize elabo-
ration. An independently used absolute clause construes a sequence together with
a relevant finite clause. The question is whether the number of absolute clauses
used independently is increasing or decreasing and whether the function distri-
bution of the independently used absolute clauses is consistent with that of the
dependently used absolute clauses (see Table 7.2).
To analyze the historical distribution of independently used absolute clauses,
we will convert the frequency in each phase into the standard frequency of per
hundred million words (see Fig. 7.1).
Figure 7.1 shows that the historical distribution of the independently used
absolute clauses is irregular and shows no clear tendency of increase or decrease.
The number of absolute clauses used dependently began to increase in 1910s and
increased rapidly after 1980s. Comparatively, the distance between the two distri-
bution lines became broader gradually after 1910s and broadened out rapidly after
1970s. Therefore, under the condition that the total number of absolute clauses is
on the increase, absolute clauses used independently are basically unchanged in
number, indicating that the number of the latter is relatively on the decrease.
According to the clause complex theory in SFL, independently used clauses
tend to be paratactic. Absolute clauses used independently are paratactic no matter
4000
3500
3000
2500
2000
1500
1000
500
0
1810s 1820s 1830s 1840s 1850s 1860s 1870s 1880s 1890s 1900s 1910s 1920s 1930s 1940s 1950s 1960s 1970s 1980s 1990s 2000s
Independent 169 58 87 100 261 381 210 113 68 172 79 285 142 197 212 259 575 292 465 379
Dependent 1016 1126 1590 1371 2143 1917 1665 1437 1573 1484 1247 1887 1902 1976 1988 2582 2847 2528 3293 3565
7.2 Case
Statistics shows that there are only a small number of absolute clauses with
personal pronoun subject. Table 7.3 displays the quantity of the case-marked
absolute clauses with personal pronoun subject retrieved from the three corpora
based on relevant regular expressions.
In the three corpora, the number of case-marked personal pronoun subjects
accounts for 3.34 % of the total of absolute clauses and is uniformly distributed.
The propositions in the three corpora are very close. However, the propositions
of accusative pronouns to nominative pronouns in the three corpora are quite
different (see Fig. 7.2).
According to the previous analysis, this is mainly because before the middle
of the twentieth century, absolute clauses gave priority to nominative personal
pronoun subjects in COHA, with quite few accusative pronouns. After 1960s,
accusative pronouns began to increase rapidly in number.
The reason why there are so few absolute clauses with personal pronoun
subject is that the subjects of absolute clauses are not co-referential with those of
the primary clauses and are less dependent on the subjects of the primary clauses.
The purpose of using absolute clauses is to avoid the subjects of the two clauses
referring to the same person or thing (Onions 1905 [2010]). Personal pronouns
do not carry semantic content and they usually refer back to an antecedent in
50
45
40
35
30
25
20
15
10
5
0
Brown BNC COHA
Accusative 38.46 46.77 21.11
the precedent context. For example, of the 28 absolute clauses with nominative
personal pronoun subject retrieved from the Brown Family Corpora, there are five
he and five she, for example
7-7a. There they continued their studies at the university, she in art, he in archi-
tecture (BROWN_A).
b. At the end, Ruth and the Indian have exchanged cultural identities, she with
an Indian name and he, a Jewish one (FROWN _G).
The subjects of such absolute clauses are co-referential with those of the
p rimary clauses and they form a contrast. Neither of the personal pronoun sub-
jects is completely co-referential with that of the primary clause. In 7-7a, the per-
sonal pronouns also imply that the subject they of the primary clause contains a
male and a female, and in 7-7b, the he provides the information that the Indian in
the primary clause is a male. Certainly, these two personal pronouns can also be
reduced to the corresponding nouns, failing to avoid simple lexical repetition.
7-8a. The wife, Amra, and her lover are both savagely portrayed, she as incarnate
sensuality, “voluptuous” and “indolent”, possibly “a mischief maker”…
(BROWN_G).
b. And so they went, he choosing of all places an inn near Medmenham Abbey
(BROWN_K).
Although she in 7-8a and he in 7-8b do not form a contrast with other personal
pronouns, both are not co-referential with the subject of the primary clause, and
both convey some new information. For example, in 7-8a, she forms a contrast
with a part of the subject, her lover, so it can be recovered to the wife, Amra, and
in 7-8b, he implies that there is a male in they.
7-9a. The big brains—they more than doubled in size from Lucy’s—did not
appear until about 2 million years ago (FROWN_F).
b. We as Black people, we can control what we do to each other (Clob_G).
they in 7-9a and we in 7-9b are co-referential with the subjects of the primary
clauses. However, such absolute clauses are not grammatically acceptable. If the
subjects are removed, the former will become a normal non-finite clause and the
latter a prepositional phrase.
7-10 We’ll leave the car and work down five or six houses on foot, me on one
side, you on the other (CROWN_L).
Like 7-7, the subjects of the two absolute clauses in 7-10 refer back to the subject
of the primary clause, but the accusative subject me appears. Although you is not
case-marked, it can also be inferred as an accusative. Even if not, nominative and
accusative personal pronouns can also be used together sometimes, for example,
7-11 Both of us had gone out to Africa, shortly after Oxbridge, at the age of
25—he to edit Drum magazine, me to edit the Nyasaland Times (Clob_G).
150 7 Discussions
It can be seen that the case choice of personal pronoun subjects of absolute clauses
is arbitrary when there is a contrast. If there is not an antecedent in the precedent
context nor a contrast is formed, accusative personal pronouns will be first consid-
ered, for example,
7-12a. Me being a less classy guy than the Pope, I wolf whistled (Clob_N).
b. …every course of action…will result in the economy collapsing like a house
of cards in a hurricane, us all losing our jobs and homes… (Clob_G).
The main function of reflexive pronouns is to emphasize. Therefore, it is not
arbitrary to choose reflexive pronouns which cannot be replaced by nominative pro-
nouns. In the following example, if himself is changed into he, it will be co-referen-
tial with the subject of the primary clause, which is not grammatically acceptable.
7-13 As editor of the Criterion, Eliot at this period often used anthropological
material, himself selecting books for review and reading every word of what
would appear in print (BNC_ACAD).
Since reflexive pronouns always have antecedents, reflexive pronoun subjects
can be omitted when the function of emphasis is not foregrounded, and hence
form non-finite clauses, verbless clauses, or nominal groups.
7-14a. To this day, Hardy, himself having achieved the double of classical and
popular success as an actor, speaks of him at that time with unaffected
adulation (BNC_MISC).
b. After they had deposited their bags at the hotel, itself ramshackle and run-
down, they had gone on to the hospital (BNC_FIC).
c. The mother-of-pearl shimmers in the background of a Henry McFee, itself
a formula derived from Renoir (BROWN_J).
Sometimes, however, the removal of reflexive pronouns may lead to the change
or ambiguity of the logico-semantic relations, for example,
7-15a. “I suppose he does,” said Joan vaguely, herself nurturing a secret fondness
for that prince (BNC_FIC).
b. I made a rule that all gifts should be submitted for approval to the archi-
tect, himself a fine artist (LOB_G).
c. According to William Julius Wilson, of the University of Chicago, himself
black, poor blacks are even more isolated because any rich blacks who can
get out do so (BNC_MAG).
d. Now, his pupil and successor, Marcel Dupre, himself in his seventies and
a pioneer of organ records, has re-recorded it there in a coupling with
Widor’s fifth and “Gothic” symphonies (LOB_C).
Both in 7-15a and in 7-15b, the absolute clauses are adverbial clauses of cause.
If the reflexive pronouns are removed, they will be changed into a non-finite
clause of extension and a nominal group of appositive. Both in 7-15c and in 7-15d,
the absolute clauses are adverbial clauses of concession. If the reflexive pronouns
are removed, they will be changed into verbless clauses of elaboration.
7.2 Case 151
7.3 Style
Observations based on the Brown Family Corpora and the BNC show that the sty-
listic distributions of absolute clauses are significantly different. Absolute clauses
tend to occur in the style of fiction, but are seldom used in learned or spoken styles.
152 7 Discussions
This is not in agreement with the traditional view that absolute clauses are formal
and infrequent.
Why are absolute clauses popularly used in fiction? This may be due to the fact
that fiction is characterized with dynamic narration and static description. This
characteristic can also be reflected in the choice of language form. The first sen-
tence of the English novel Pride and Prejudice is famous not only in its semantic
content but also in its particular language form which will in turn glorify its basic
semantic content. Absolute clauses as a substitute form of finite clauses can func-
tion as a means to foreground the semantic content.
According to SFL, absolute clauses have their transitivity structure, thematic
structure, and information structure, but no mood structure. The interpersonal
metafunction of language is the function to enact social processes through mean-
ing. In narration or description, a social process is enacted, and then, narration or
description continues. Absolute clauses are used to avoid the frequent enacts of
social processes.
When used in the paratactic relations, absolute clauses of extension add new
content to the primary clauses. The corresponding finite forms of such absolute
clauses are the continuing clauses in paratactic clause complexes. Absolute clauses
of extension and enhancement are all dependent clauses and hence are flexible
in position. This facilitates to highlight the propositions enacted by the primary
clauses. Without the use of absolute clauses, the primary information will be
submerged into various secondary information, for example,
7-17 Billee wagged his tail appeasingly, turned to run when he saw that appease-
ment was of no avail, and cried (still appeasingly) when Spitz’s sharp teeth
scored his flank. But no matter how Spitz circled, Joe whirled around on his
heels to face him, mane bristling, ears laid back, lips writhing and snarling,
jaws clipping together as fast as he could snap, and eyes diabolically gleam-
ing—the incarnation of belligerent fear. So terrible was his appearance that
Spitz was forced to forego disciplining him; but to cover his own discomfi-
ture he turned upon the inoffensive and wailing Billee and drove him to the
confines of the camp (COHA_FIC).
As can be expected, the typical characteristics of learned texts are the wide use
of passive voices and nominalizations, which are effective methods to objectivize
the proposition. Nominalization is also a method to compact the information con-
strued by a clause into a nominal group. Because of the compact of information,
clause complexes are not encouraged to use in learned, nor are absolute clauses.
The small number of absolute clauses in learned texts realize hypotactic elabora-
tion, for the purpose of the learned texts is not to narrate or describe but to explain
or argue, for example,
7-18a. Five papers are now in circulation, two from consume groups, one from
managers, and two from medical organisations (BNC_ACAD).
b. The test subjects are shown three video clips involving the same simple
action: a hand grasping a teacup (CROWN_J).
7.3 Style 153
In 7-18a, the three absolute clauses are the explanations to Five papers.
In 7-18b, the absolute clause is used to explain the same simple action.
According to the data, absolute clauses of elaboration do not take the advan-
tage in learned over the other styles. This is because the subjects of most absolute
clauses of elaboration are personal pronouns, and personal pronoun subjects are to
be avoided as much as possible in learned texts. For example, of the 61 absolute
clauses with personal pronoun subject retrieved from the Brown Family Corpora,
only two inanimate third person singular reflexive pronouns itself occur in learned.
Therefore, absolute clauses of elaboration, even though small in number, are still
mostly appear in other styles than in learned, for example,
7-19a. Here’s something about Miss Leefolt: she not just frowning all the time,
she skinny (CROWN_K).
b. It looks real nice, them sitting there, just the two of them (BNC_FIC).
c. It was still there, they both thought, she with greed and apprehension, he
with alarm (BNC_FIC).
In learned, the subjects of absolute clauses of elaboration are mostly common
nouns or pronouns. Of the 32 subjects of absolute clauses of elaboration, there are
only three reflexive personal pronouns (itself) retrieved from academic in the BNC,
for example,
7-20
On 26 December the upper house of the USSR Supreme Soviet, itself
inquorate, voted a formal end to the original treaty of union (BNC_ACAD).
7-21a. Three were misdiagnosed as having colitis; one in childhood, and two oth-
ers at age 38 and 42 years respectively (BNC_ACAD).
b. Play another sound effect—perhaps a door opening, keys jangling, a car
moving off, a bomb exploding, or a dog barking—and it becomes difficult
not to link the two sounds together and make them part of the same story
(BNC_ACAD).
7.4 Time
The COHA-based research shows that over a span of 200 years, the historical dis-
tribution of absolute clauses manifests a significant difference, and different func-
tions of absolute clauses have different evolution trends. Extension is increasing;
enhancement is gradually decreasing but keeps steady in recent dozens of years
(not dying out at all); elaboration shows no obvious historical changes.
Although traditional grammar (e.g., Quirk et al. 1985) insists that absolute
clauses are not frequently used in today’s English, some fixed expressions such
as Weather permitting, All things considered, There being… and It being… are
still popularly seen. Most of such stereotyped absolute clauses realize hypotactic
enhancement. Table 7.4 shows some of the most frequently occurring absolute
154 7 Discussions
clauses of enhancement retrieved from the three corpora. Here, the frequencies
of absolute clauses will not be compared, because different regular expressions
have been used in different corpora. For example, the regular expressions can be
used to retrieve absolute clauses beginning with nouns or pronouns in the BNC,
and absolute clauses composed of nouns or pronouns and present participles.
Therefore, such constructions as all things considered, there being… can be
retrieved from neither of the two corpora.
We will now take the fixed expressions (the) weather permitting/being…
retrieved from COHA as examples to analyze their historical distributions. The
data are shown in Table 7.5.
When combining the two expressions with the together and combining the
other two without the together, we have 57 instances with the and without the
each. To facilitate comparison, we will convert the frequencies with the and those
without the in the 20 phases into the standard frequencies of per hundred million
words (see Fig. 7.3).
Figure 7.3 shows that although the two lines are not regularly distributed, they
manifest a general trend: weather permitting/being increases and the weather per-
mitting/being decreases. The reason is that weather permitting/being has been
grammaticalized into a fixed expression, realizing the relation of condition. The
formal indicator of grammaticalization is the loss of the article the. The con-
struction the weather permitting/being has not been grammaticalized into a fixed
expression because there exists the article the in form on the one hand, and on the
other hand, it can realize different relationships. In the following, we will analyze
the trend of grammaticalization of the absolute clauses we collected from COHA
by reference to the articles the and a (Fig. 7.4).
An important type of grammaticalization of absolute clauses is concerned
with prepositions, among which the construction “noun + preposition + noun”
can be taken as a typical example pattern. Absolute clauses of attendant circum-
stance such as hat in hand are compressed semi-idiomatic absolute constructions,
and they can be expanded into augmented absolute constructions (e.g., with a hat
in his hand) (Jespersen 1949). Based on this explanation, such phrases as hand
7.4 Time 155
70
weather permitting/being the weather permitting/being
60
50
40
30
20
10
0
1810s 1820s 1830s 1840s 1850s 1860s 1870s 1880s 1890s 1900s 1910s 1920s 1930s 1940s 1950s 1960s 1970s 1980s 1990s 2000s
70
60
50
40
30
20
10
0
1810s 1820s 1830s 1840s 1850s 1860s 1870s 1880s 1890s 1900s 1910s 1920s 1930s 1940s 1950s 1960s 1970s 1980s 1990s 2000s
Fig. 7.4 Historical distribution of absolute clauses with and without article in COHA (per hundred
million words)
3500
3000
2500
2000
1500
1000
500
0
1810s 1820s 1830s 1840s 1850s 1860s 1870s 1880s 1890s 1900s 1910s 1920s 1930s 1940s 1950s 1960s 1970s 1980s 1990s 2000s
Extension 2116 1516 2497 2823 2823 3014 3028 2668 2524 2751 2568 2830 2227 2403 2587 2081 2431 2216 2602 2645
7.5 Summary
References
Huang, G.-W. (1998). A functional analysis of the English causative structure. Journal of Foreign
Languages, 1, 12–16.
Jespersen, O. (1949). A modern English grammar on historical principles. London: Allen & Unwin.
Onions, C. T. (1905 [2010]). An advanced English syntax. Whitefish: Nabu Press.
Quirk, R., Greenbaum, S., Leech, G., & Svartvik, J. (1985). A comprehensive grammar of the
English language. London, New York: Longman.
Wheelock, F. M., & LaFleur, R. A. (2005). Wheelock’s latin (6th ed.). New York: HarperCollins.
Chapter 8
Conclusion
The findings of this research can be summarized by the answers to the questions
proposed in Chap. 4.
What types of relation can absolute clauses realize?
Absolute clauses as non-finite clauses with subject have the same meaning
potential as all non-finite clauses. However, not all non-finite clauses with sub-
ject can form absolute clauses. Absolute clauses defined by traditional grammar
can be classified into two syntactic types: clausal adjuncts and attendant circum-
stances. Relevant research proposes a third type: appositives. According to the
relation system of clause complex in SFL, the three types of absolute clauses cor-
respond to the dependent clauses of extension, enhancement, and elaboration in
clause complexes. Since absolute clauses can realize hypotactic expansion, they
have the potential to realize hypotactic projection. According to the principle of
double transitivity analysis, absolute clauses can be seen as embedded non-finite
clauses realizing circumstances and participants. In addition, the corpus-based
research shows that besides realizing hypotaxis, there are also a large number of
absolute clauses standing alone, i.e., they are used as independent sentences. In
spite of being independent, absolute clauses still construe a relation of elabora-
tion, extension, and enhancement with the preceding or following sentences. This
kind of relation is cohesive but not structural. Further analysis shows that abso-
lute clauses can realize not only hypotaxis but also parataxis. Four types of abso-
lute clauses can be recognized in structure: absolute appositive clauses, absolute
adjunct clauses, absolute subject clauses, and absolute complement clauses.
Different types of absolute clauses are different in independence. The analysis
of independence of absolute clauses from two dimensions, the independent ten-
dency of absolute clauses and that of primary clauses, show that different types of
absolute clauses are on a cline of independence. The independently used absolute
clauses are the most independent, absolute clauses in paratactic clause complexes
are the second most independent, and then absolute clauses in hypotactic clause
complexes, the least independent. In hypotaxis, absolute clauses of extension are
more independent than those of enhancement. Absolute of participants are the
least independent because they are restricted by the main verbs of the primary
clauses. Seen from the relation between the subjects of absolute clauses and the
main verbs, absolute clauses of subject are more independent than those of com-
plement. But seen from the requirement of the whole construction by the main
verb, absolute clauses of complement are more independent than those of subject.
The general principle is that the closer the relation between the absolute clauses
and the main verbs is, the less independent the absolute clauses are, or vice versa.
What are the synchronic and diachronic distributions of absolute clauses?
To answer this question, we extracted from the Brown Family Corpora,
the BNC, and COHA 10,930 absolute clauses of expansion and 47,889 absolute
clauses of projection. The quantitative diachronic and synchronic research on
these absolute clauses shows that the number of occurrences of absolute clauses
in the recent 200 years is not decreasing but increasing. This is mainly reflected in
the obvious increase of absolute clauses of extension. The traditional opinion that
absolute clauses are decreasing can only explain the tendency of absolute clauses
of enhancement. Corpus data do not show that absolute clauses of enhancement
are diminishing. The data of the most recent 50 years show that the distribution of
absolute clauses of enhancement tends to be stable. Analysis shows that in the pro-
cess of historical development, absolute clauses have been being grammaticalized.
Some absolute clauses of enhancement have been fixed into stereotyped expres-
sions, realizing some certain functions, hence are impossible to disappear.
The observation based on the Brown Family Corpora shows that absolute
clauses are mostly in fiction and are seldom in learned. This is obviously not in
agreement with traditional grammar according to which absolute clauses are popu-
lar in formal texts, because the register of learned is the most formal, with that of
press, general prose, and fiction following. However, the distribution of absolute
clauses in these four registers is exactly the other way round, respectively, fiction,
general prose, press, and learned. There are no spoken data in the Brown Family
Corpora. The BNC-based research shows that absolute clauses are rarely seen in
either the informal spoken texts or the formal learned texts and still occur in fic-
tions. Therefore, the distribution of absolute clauses is different in register, but this
difference has nothing to do with whether the register is formal or not. The rea-
son why absolute clauses are mostly used in fiction is that there are narrations of
events and descriptions of scenes. Absolute clauses turn out to be equipped with
such functions.
In addition, we have carried out a research based on the Brown Family Corpora
on the case choice of absolute clauses discussed in traditional grammar. Of the 61
absolute clauses with personal pronoun subject, most are concentrated in fiction
and general prose, and fiction dominates. This is mainly because the total num-
ber of words of the register of fiction is larger. When transferred into the stand-
ard frequency of per million words, the frequency in general prose is obviously
8.1 Main Findings of This Research 161
higher than that in fiction, even the average frequency in fiction is slightly lower
than the total average frequency. The study of the case distribution on the 49 case
marked absolute clauses with personal pronoun subject shows that the personal
pronoun subject can be either nominative or accusative and the distribution of case
in the recent 50 years does not show any significant diachronic or regional differ-
ences. However, such a small number of absolute clauses with personal pronoun
subject show significant register differences. In all the four registers, the number
of nominative pronouns is clearly smaller than that of accusative pronouns. The
difference between the nominative pronouns and the accusative pronouns is the
largest in general prose, and the proportion in fiction is nearly at the total aver-
age level. Only in press are there more nominative pronoun subjects than accusa-
tive pronoun subjects, and the number of nominative pronoun subjects is larger
than that in fiction. Personal pronouns do not tend to occur in the learned regis-
ter, neither do absolute clauses with personal pronoun subjects. In the six corpora,
there are only two absolute clauses with itself as subject. Data show that when the
subject of an absolute clause has to be a personal pronoun, people do not tend to
choose nominative pronoun subject, rather they tend to choose accusative pronoun
subject. The COHA-based research presents the same result. The case of personal
pronoun subjects of absolute clauses shows obvious diachronic difference in the
span of 200 years. Although the total number of nominative pronoun subjects is
much larger than that of accusative pronoun subjects, the distribution of personal
pronoun subjects shows that the number of nominative pronouns has been decreas-
ing and that of accusative pronouns increasing since the beginning of the twentieth
century. In addition, the number of accusative pronoun subjects began to surpass
that of nominative personal pronoun subjects in the 1980s phase.
8.2.1 Limitations
The corpus-based systemic functional research of absolute clauses has the follow-
ing problems:
First, there is no way to extract all absolute clauses from the corpus through
automatic retrieval and manual extraction.
The advantage of computer processing lies in its being able to process more
data quickly and to reveal some implicit language features. However, there is
still a considerable distance between computer processing and manual process-
ing in depth, accuracy, flexibility, richness, etc. “Automatic analysis gets harder
the higher up we move along the hierarchy of stratification” (Halliday and
Matthiessen 2004: 49). That is to say, the higher the grammatical rank is, the more
difficult the automatic analysis will be. For example, automatic analyses can deal
with any models described in words and models of lower lexico-grammatical
ranks, but cannot conduct systemic functional analyses of clauses completely or
162 8 Conclusion
8.2.2 Further Research
“As Jones said, ‘a science without difficulties is not a science at all’” (Halliday and
Matthiessen 2004: 35). Many scholars advocate seeking a balance “between vol-
ume of analysis and richness of analysis: low-level analysis can be automated to
handle large volumes of text, but high-level analysis has to be carried out by hand
for small samples of text” (ibid.: 49). Absolute clauses as non-finite clauses are at
a higher level in rank and so are suitable for manual processing. This requires that
in the future research, a special attention be paid to the intercrossing between con-
cepts and the complementarity of methods: complementarity between quantitative
8.2 Limitations and Further Research 163
References
CC Coordinating conjunction
CD Cardinal number
DT Determiner
EX Existential there
FW Foreign word
IN Preposition or subordinating conjunction
JJ\w* Any adjective
JJ Adjective
JJR Adjective, comparative
JJS Adjective, superlative
LS List item marker
MD Modal
N\w+ Any noun
NN Noun, singular or mass
NNS Noun, plural
NP Proper noun, singular
NPS Proper noun, plural
PDT Predeterminer
POS Possessive ending
PP Personal pronoun
PP$ Possessive pronoun
RB\w* Any adverb
RB Adverb
RBR Adverb, comparative
RBS Adverb, superlative
RP Particle
SYM Symbol
TO To
UH Interjection
VB\w* Any form of the “be” verb
See Tables B.1, B.2, B.3, B.4, B.5, B.6, B.7, B.8 and B.9.
Fig. C.2 Example retrievals of personal pronoun + nonfinite element constructions from the Frown
Corpus using the regular expression: \S+_(,|SENT|:)\s\S+_PP(\S+_RB\w*)*\s(\S+_TO\s\
S+_V[BDHV]|\S+_(V(B|D|H|V)[GN]|JJ\w*|IN|DT|N\w+)
178 Appendix C: Corpus Retrieving Demonstration
Fig. C.3 Example retrievals of noun or pronoun + present participle constructions from BNC
using the regular expression: [y*] [nn*]|[p*] [v?g*] (0-2) [y*]