Compacting the Message: the Use (Or Non-Use) of Participle Clauses in Native and Non-Native

Compacting the Message: the Use (Or Non-Use) of Participle Clauses in Native and Non-Native

1

In: Aarts, J., de Mönnink, I. and Wekker, H. (eds) (1997) Studies in English Language and Teaching. Rodopi: AmsterdamAtlanta, pp. 185-198.

ON IDENTIFYING THE SYNTACTIC AND DISCOURSE FEATURES OF PARTICIPLE CLAUSES IN ACADEMIC ENGLISH: NATIVE AND NON-NATIVE WRITERS COMPARED

Sylviane Granger

Centre for English Corpus Linguistics

Université Catholique de Louvain

1. Non-finite clauses in EFL grammars

In comparison to the detailed treatment which finite clauses receive in EFL grammars, the description of non-finite clauses appears very perfunctory. Not only are they insufficiently described, they are also usually scattered across several sections of the grammar, thus making it very difficult for EFL learners to form an overall picture of the part they play in discourse. In a form-based grammar such as Thomson & Martinet (1986), non-finites are dealt with in sections on participles, infinitives, gerunds, relative pronouns and clauses, etc., while in a function-based grammar like Downing & Locke (1992) they are found in sections as diverse as ‘the development of the message’, ‘expressing intrinsic features of things’, ‘enhancing the message’, etc. In most cases, non-finites are presented as free variants of finites, with no mention of any factors favouring one type over the other. It is not rare to find statements such as the following: “The non-finite verb forms -ing, -to infinitive and -en participle are used non-restrictively to express the same meanings as the finite forms” (Downing & Locke 1992:286). Two notable exceptions are Close (1975) and Johansson & Lysvag (1986). Close presents all the uses of non-finite clauses in a special section entitled ‘Condensing the sentence’, together with other abbreviatory/compacting/synthetic devices, such as verbless clauses, complex prepositional phrases, ellipsis, pro-forms, etc., while in their excellent two-volume grammar, Johansson & Lysvag give a very detailed description of these structures in two chapters, one on non-finite expansion of the sentence and the other on -ing constructions.

There are two related reasons why non-finite clauses should receive more prominence in EFL grammars, especially those aimed at advanced learners. Firstly, non-finites and in particular participle clauses - the topic of this article - have been found to figure prominently in some registers of English, notably narrative and academic writing. Beaman (1984:66), for instance, shows that finite subordinate clauses are more common in spoken narratives and nonfinite clauses in written narratives. Chafe & Danielewicz (1987:102), on the other hand, show that high frequency of participles is a distinguishing feature of academic writing. They compare four registers of English - two spoken and two written - and conclude that “language other than academic writing makes considerably less use of participles”. As the improvement of stylistic proficiency is an important objective at an advanced stage of learning, the stylistic features of non-finite clauses ought to receive more attention. The second reason for giving more prominence to non-finite clauses is that they are a major means of syntactic compression (Greenbaum 1988). The finite adverbial clause in the sentence As he was of noble extraction, he found it easy to mix with the local aristocracy can be abbreviated to a nonfinite clause (Being of noble extraction,he found it easy...) or further compressed into a verbless clause (Of noble extraction, he found it easy...). A better mastery of non-finites could therefore help learners to develop a more compact, integrated style.

2. Participle clauses

The category of non-finite clauses clauses, ie clauses whose verb element is non-finite (Quirk et al, 1985:992), consists of participle clauses and infinitive clauses. Sentences (1) to (3) illustrate the three major syntactic functions of participle clauses: nominal in (1), adverbial in (2) and postmodifying (or: adnominal) in (3).

(1) I enjoyed spending my holidays with them.

(2) As mentioned in section 1, several statistical tests were used..

(3) People working in this field have made considerable progress.

Nominal clauses contain -ing participles functioning as subject, object, subject complement, etc. Unlike for the other two types, there are no systematic finite alternatives. In sentence (1), for instance, the -ing participle is triggered by the verb avoid and has no finite counterpart. This makes nominal clauses less interesting to investigate from a stylistic point of view and they have therefore been excluded from the current investigation.

Adverbial -ing and -ed clauses express a variety of semantic relationships, which can be expressed by means of a subordinator - a conjunction (ex. 4) or a preposition (ex. 5) - or left implicit (ex. 6-8). They usually have no subject of their own but when they do, as in sentence (8), they are referred to as ‘absolute clauses’. Adverbial clauses which are not introduced by a subordinator are called ‘supplementive clauses’ by Quirk et al (1985:1123). As their adverbial status is often unclear (see section 6), I have classified them in a category of their own, reserving the term ‘adverbial’ for subordinator-headed clauses.

Adnominal participle clauses, which function as postmodifiers to nouns or pronouns, are usually referred to as ‘reduced relative clauses’ (ex. 9 & 10). Indeed, they can be viewed as shorter versions of full relative clauses (A report that was written by my colleague; a taste which resembles that of soapy water).

(4) When returning the merchandise, be sure to bring your receipt.

(5) On becoming a member, you will receive a membership card and a badge.

(6) Returning to my village after thirty years, I met an old schoolteacher.

(7) Stated bluntly, he had no chance of winning.

(8) No further discussion arising, the meeting was brought to a close.

(9) A report written by my colleague appeared last week.

(10) This is a liquid with a taste resembling that of soapy water.

(from Quirk et al 1985)

Adverbial and supplementive clauses bear the danger of ‘dangling’ or ‘unattached’ participles, which are condemned in all grammars and writing textbooks, for native and non-native speakers alike. According to Greenbaum (1991:159) “a dangling modifier has no subject of its own, and its implied subject cannot be identified with the subject of the sentence though it can usually be identified with some other phrase in the sentence”. To illustrate this phenomenon, he juxtaposes erroneous sentences and their corrected versions (cf. ex. 11 & 12).

(11) dangling Being blind, a dog guided her across the street.

correctedBeing blind, she was guided across the street by a dog.

(12)danglingAfter turning the radio off, the interior of the car became silent.

correctedAfter she (or I, etc.) turned the radio off, the interior of the car became

silent.

In this article I investigate the use of participle clauses in the academic writing of native English students and advanced EFL learners and compare the results in terms of frequency of occurrence, syntactic pattern and discourse function. The analysis is based on two 45,000 word corpora extracted from the ICLE (International Corpus of Learner English) and LOCNESS (Louvain Corpus of Native English Essays) databases (for more information, see Granger forthcoming). The non-native (NNS) corpus consists of argumentative essays written by advanced EFL learners and is subdivided into three 15,000 word subcorpora - NNSF, NNSS and NNSD - which cover writing by French, Swedish and Dutch students respectively. The native speaker (NS) corpus covers similar writing by American English students; for comparison purposes, it has also been split into three similar-sized chunks - US1, US2 and US3.

3. Overall frequency of occurrence of participle clauses

Table 1 gives the overall frequency of occurrence of participle clauses in the NS and NNS corpora. It brings out a statistically significant underuse of these clauses - both -ed and -ing - by EFL learners (p<0.005). Indeed, there are half as many participle clauses in the learner as in the native speaker corpus. The two corpora are similar, however, in the distribution of -ing and -ed participles: two thirds -ing and one third -ed.

Table 1: Overall frequency of occurrence of participle clauses in NS and NNS corpora

NS / NNS
Ved / 165 (34.5%) / 96 (39.5%)
Ving / 310 (65.5%) / 147 (60.5%)
Total / 475 (100%) / 243 (100%)
Nr words / 45,600 / 46,211

In order to check the reliability of these figures, it is useful to look at the frequency of participle clauses in the six subcorpora. The results are given in Table 2. The figures printed in bold in the table represent statistically significant differences within the NS and NNS corpus.

Table 2: Overall frequency of occurrence of participle clauses in NS and NNS corpora: subcorpus breakdown

US1 / US2 / US3 / NNSF / NNSS / NNSD
Ved / 75 / 50 / 40 / 40 / 36 / 20
Ving / 119 / 96 / 95 / 44 / 53 / 50
Total / 194 / 146 / 135 / 84 / 89 / 70
Nr words / 15,030 / 15,666 / 14,904 / 15,246 / 15,810 / 15,155

The overall underuse brought out by Table 1 is confirmed here: all three learner varieties prove to underuse participle clauses significantly (p<0.005), with frequencies ranging from 70 to 84, as opposed to 135 to 194 for the NS subcorpora. However, while there are no significant frequency differences among the NNS subcorpora, the NS corpus displays some variability: US1 contains a significantly higher number of participle clauses than the other two NS subcorpora.

A look at the respective distributions of -ed and -ing shows that the latter is much more stable across subcorpora than the former: it is consistently lower in the learner corpus, with frequencies ranging from 44 to 53 (as opposed to 95 to 119 in the NS corpus). The frequency of -ed structures, on the other hand, is highly variable in both the NS and NNS corpora with two subcorpora standing out from the others: US1 displays significantly more -ed structures than US2 and US3, while the Dutch learner corpus (NNSD) contains a significantly lower number of these structures than either the French or Swedish corpora.

There are several possible reasons for this underuse of participle clauses by the EFL writers may be explained by several factors. First there is the fact that they receive little attention in EFL grammars and learners will thus be unsure as to how and when they should use them. In addition, they may consciously or unconsciously avoid using them because of the spectre of the dangling participle. If this is true, it is a shame, as learners already have to fight against a natural tendency to overtextualize, to use more words than is necessary. Finally, transfer from the mother tongue may also play a part. According to Kortmann (1995:229) English “stands out among the Germanic languages with respect to its predilection for the use of adverbial participles”.(1) And in fact when one attempts to translate the English participle clause types into the other three languages under investigation, ie French, Swedish and Dutch, several of the English structures turn out to be either inexistent or much less frequently used. (2) It is interesting to note that this is especially true of -ing clauses, which have proved to be the most consistently underused type. It is therefore possible - indeed likely - that learners underuse some of the participle structures because they have a more restrictive use in their mother tongue. However, in the absence of reliable text-based crosslinguistic descriptions of these structures, one should be careful not to draw hasty interlingual conclusions.

Overall frequencies of participle clauses may well hide significant differences in the use of individual clause types. The following sections present a detailed description of each of the three major syntactic types of participle clause - adnominal, adverbial and supplementive - in the learner and native speaker corpora.

4. Adnominal participle clauses

As Table 3 shows, the category of adnominal clauses is the largest category of participle clauses: it accounts for approximately half the number of occurrences in the two corpora, with -ed postmodification being the dominant tendency.

Table 3: Frequency of occurrence of adnominal clauses in NS and NNS corpora

NS / NNS
Adnominal / 226 (47.5%) / 127 (52%)
Ved / 133 / 79
Ving / 93 / 48
Total / 475 (100%) / 243 (100%)

Adnominal clauses are significantly underused by learners, with nearly twice as many occurrences of both -ed and -ing participles in the NS corpus. The frequencies of -ed and -ing per 1,000 words are respectively 1.7 and 1.0 in the learner corpus, as opposed to 2.9 and 2.0 in the native speaker corpus. (3) The subcorpus breakdown of adnominal clauses reveals that the frequencies for -ing clauses are consistently low, while the underuse of -ed is mainly due to a very low frequency in the Dutch learner corpus. The reason accounting for this underuse does not seem to be transfer-related as past participial relative clauses exist in Dutch as well as in French and Swedish. The consistent underuse of -ing clauses, however, may well be partly due to the fact that these structures are either impossible or comparatively much less frequent in the learners’ mother tongues.

Sentences (13-20) illustrate how adnominal clauses are used in the NS corpus.

(13) Obviously, the behaviour demonstrated by the children in the accounts was not typical for that child, or really any child for that matter.

(14) Further evidence of this is demonstrated by a 1935 study conducted by Robert Dann.

(15) The thought conveyed in this prayer is shallow in meaning.

(16) All the things presented here could make for an effective argument against continuing genetic research.

(17) it will be ultimately decided that there should be a ban on gays serving in the military.

(18) no evidence was presented to suggest that reports using notebooks and pencils posed any security risks.

(19) It could easily become a solution for people suffering from severe depression or those diagnosed with HIV or AIDS, who want to save themselves from the pain that they will later endure.

(20) The pain experienced by someone dying of cancer is enormous.

It is not my intention to provide a detailed account of the syntactic and discourse functions of reduced relative clauses at this point, especially in view of the dearth of information in English grammars. However, the corpus study shows that these clauses play an important role in academic discourse, both in rhematic position, where their use is governed by the principles of end-focus and end-weight (cf. ex. 14, 17 & 19), and thematic position, where their anaphoric function is brought out by the presence of back-referencing elements such as definite noun phrases (the children/the accounts in 13), demonstrative pronouns (this in 14) or text-referring adverbs (here in 16). It is also interesting to note that there is a high occurrence of agentful clauses in the non-finite clauses. Approximately 40 % of the past participle clauses contain a by-agent, a proportion which is markedly higher than that found for finite passive forms, where it rarely exceeds 20% (see Svartvik 1966).

It has been shown that the EFL writers underuse adnominal clauses. In fact, there is also a reasonable degree of misuse. In several sentences, such as (21-25), the choice of a reduced relative clause has been found infelicitous by the native speaker corrector and replaced by a finite relative clause (4).

(21) Not to mention these ethnical minorities not feeling at home. (correction: who don’t feel at home)

(22) All the countries making part of that Europe will live hand in hand. (correction: who will make up Europe)

(23) many Europeans fear that they will be the eyewitnesses of the birth of a nation leaving the citizens aside. (correction: which will leave)

(24) Many dreams and imaginations of the people living in the early part of the century are now a reality. (improvement: who lived)

(25) ...to fight starvation and diseases endangering the lives of the natives. (improvement: which endanger)

Nonfinite clauses are obviously not free variants of finite clauses. It is often difficult to say why one is chosen above the other but there seem to be a number of factors involved such as the need to remove ambiguity, the informational load of the relative clause, and so on. What is certain is that there is a need for detailed comparisons of finite and nonfinite clause types.

5. Adverbial participle clauses

Adverbial participle clauses can express a variety of circumstancial relationships - time, cause, manner, condition, etc. - which are made explicit by a conjunction or a preposition. They account for approximately one third of the clauses in the data. As shown by the figures in Table 4, it is the category that is most significantly underused by learners: there are 2.5 times as many occurrences of such clauses in the NS as in the NNS corpus (182 occurrences vs 75). The frequencies of adverbial clauses are extremely stable across the subsections of each corpus (from 58 to 65 in the NS corpus and from 21 to 28 in the NNS corpus). This holds for both conjunction- and preposition-headed clauses.

Table 4: Frequency of occurrence of adverbial clauses in NS and NNS corpora

NS / NNS
Adverbial / 182 (38.5%) / 75 (31%)
Conj+ / 39 / 18
Ved / 17 / 3
Ving / 22 / 15
Prep+ / 143 / 57
Total / 475 (100%) / 243 (100%)

5.1. Conjunction-headed clauses

Only a minority of adverbial clauses (c. 25%) are introduced by a conjunction. Table 4 shows that it is mainly the -ed conjunction-headed clauses (see ex.26-31) that are underused by learners.

(26) do students better understand scientific principles when taught from a two-model approach of origins (...) or a one-model approach?

(27) There are rules to follow when carrying out affirmative action policies which can be unconstitutional if ignored.

(28) As shown previously, racism on campus is a serious issue.

(29) As stated above, when he returned the following year to reapply, he wrote down the two drugs on his application.

(30) As stated in Time, “<quote>“.

(31) When used in business, ethics involves the relationship of what is right (...).

The three conjunctions that serve to introduce past participial adverbial clauses in the corpus are as (8 occurrences), when (7) and if (2) in the corpus. In the majority of cases, the -ed clauses are initial and have a connective function, as in examples 28 & 29, or a frame-setting function (Kortmann 1995:228), as in examples 30 and 31.

Present participle adverbial clauses are introduced by when, while and whether...or. The most frequently used conjunction is when (16 occurrences in NS and 9 in NNS). Like -ed clauses, -ing clauses are found in thematic (ex. 32) and rhematic position (ex. 33), but rhematic position is more common. Using Chafe’s (1984) terminology, we could say that -ed clauses tend to have a ‘guidepost’ function, while -ing clauses function more as some kind of ‘afterthought’. It is interesting to note that whereas the learners often use the ‘afterthought’ function, they rarely use the ‘guidepost’ one.

(32) When dealing with the history of genetic research, one should bring up eugenics, which is the study and methods of improving the species genetically.

(33) This drug is used to help revive heart attack victims while simultaneously preventing brain damage from occurring.