國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
CHAPTER 5
OF-CONSTRUCTIONS BEYOND THE CONSTRUCTION LEVEL
In this chapter, we go beyond the construction to examining the different linguistic contexts in which the of -constructions occur. First, we looked at the syntactic position where each of -construction is used in a sentential context. Then we examined whether each of -construction is post-modified and the types of post-modification are used.
Finally, the above two factors were considered simultaneously with construction level factors considered earlier in Chapter 4 by means of a multivariate statistical analysis which highlights any significant associations among various factors.
The following discussion is divided into four sections. Section 5.1 presents the types of syntactic position each of -construction occupies. Section 5.2 discusses how the of -constructions are postmodified. Section 5.3 considers the statistical results of hierarchical cluster frequency analysis. Section 5.4 presents a discussion, followed by a summary in Section 5.5.
5.1 The Syntactic Positions of Of -constructions
To identify where of -constructions may be commonly found in a clause, we considered the occurrence of the of -constructions in various syntactic positions including the subject, the object, the complement, and others (e.g., non-clausal elements such as headings or parenthetical structures). Each syntactic position may be further divided into sub-types based on clause types (e.g., existential clause, copula clause). In addition to the independent clauses with the word order of subject-verb-object-complement, there are also existential clauses (there-verb-notional subject) and copular be verb constructions (subject-be verb-subject complement).
Table 5.1 presents three major positions examined in this study and provides the frequency distribution of the of -construction found in different syntactic positions. It is shown from this table that the most common type of syntactic positions where an of -construction is used is the complement (41.0%), followed by the object (30.8%) and the subject (22.6%). The examples are given below. Interestingly, the proportion
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
of the subject was about 10% less than that of the object and 20% less than that of the complement, suggesting that of -constructions do not function actively in staging the sentence (Halliday, 1985) as the role normally taken by a grammatical subject. As discussed in Langacker (1987), the subject position is the most salient focal participant as compared to other role archetypes such as the object and complement. In contrast, the less salient roles, i.e., the object and the complement, are much more likely to be taken by an of -construction. This observation is congruent with the fact that most of -constructions as the subject do not exhibit a sense of agentivity, as derived from an animate entity which is the most likely candidate for carrying out an action (e.g., Lyngfelt & Solstad, 2006). A quick search for the of -constructions which contain an animate entity at N1 in a transitive clause shows a very low proportion of 2.8 percent (30 instances out of 1,074) as shown in (5.1a).
Table 5.1 Distribution of syntactic positions of of -constructions
Types Sub-types Count Frequency Subtotal
Subject Existential 66 1.35%
Non-existential 1,074 22.00% 23.36%
Object Direct object 1,379 28.25%
Prepositional object 428 8.77% 37.02%
Complement Complement of preposition
1,420 29.09%
Subject complement 283 5.80% 34.93%
Object complement 2 0.04%
Others 229 4.69% 4.69%
Total 4,881 100.00% 100.00%
(5.1) a. Lord Taylor of Gosforth C.J. read the following judgement of the court.
(FCC-80)
b. The increasing eccentricity of its founder alienated many as well. (CS6-395)
In comparison to (5.1a), (5.1b) contains an animate entity at N2. While both instances have a transitive verb, i.e., read and alienated, (5.1a), however, appears to be higher in agentivity than (5.1b). This is because the subject of (5.1a) refers to an animate entity
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
having the ability to act on another entity but not that of (5.1b). If we examine further to look at each subtype, the majority of of -constructions can be found at the complement of preposition (29.09%), object (28.25%), and non-existential subject (22.00%), all of which accumulate to a sum of nearly 80 percent. The remaining twenty percent is comprised of positions such as prepositional object (8.77%), subject complement (5.80%), existential subject (1.35%), and others (4.69%) where sentence fragments (5.2a) and headings (5.2b) predominate.
(5.2) a. Valuing - acceptance of a value preference for a value - commitment IV (EUW-706)
b. 4.2.2 Breach of warranty (J6Y-327)
The following discussion will be divided into two parts distinguishing the copula be constructions and existential constructions from the rest. This is because these two types of constructions do not follow the simple English word order (subject-verb-object) like the majority of the concordance lines analyzed in this study. Section 5.1.1 presents the analysis of the simple clauses, and Section 5.1.2 presents the analysis of copula be constructions and existential constructions.
5.1.1 The Simple Constructions
The syntactic positions of the of -constructions found in the simple constructions with the word order of subject-verb-object are discussed in this section in both active and voices. For the active constructions, we can find the of -construction mainly at the subject, object, and complement positions. The discussion will be divided according to these three positions. Then the passive construction will be presented.
5.1.1.1 Subject
Although theof -construction found at the subject position is about ten percent less than that at the object and at the complement, the of -construction at the subject position appears to be peculiar in its own right. Quirk et al. (1985) writes that “the most typical semantic role of a subject in a clause that has a direct object is that of the
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
AGENTIVE participant: that is, the animate being instigating or causing the happening denoted by the verb” (p. 741). From previous studies on genitive alternation such as Kreyer (2003) (see Section 2.1.1), one crucial differentiation between the s-genitive and of -construction is human involvement or the degree of animate entity involved in the proposition: a low degree of human involvement has been identified to be characteristic of the of -construction. To recapitulate, a However, our corpus data shows that the of -constructions responsible for groups (e.g., a school of Dalmatian humanists) or group-membership (e.g., the Director of Education) exhibit a high degree of human involvement as illustrated in 5.3 even though the occurrence rate is as low as 4.7 percent (50 out of 1084).
(5.3) a. Some working-class members of the Legions also held open air meetings in Hackney and other locations in the East End of London at irregular intervals. (CAN-929)
b. The Court of Appeal in that case thought that the distinction between larceny by a trick and obtaining by false pretences depended on the presence in the Larceny Act of the words “without the consent of the owner,” … (FE2-635)
c. …where the Court of Chancery exercised its jurisdiction to compel a husband to make a settlement upon his wife. (ABP-485)
Example (5.3a) presents such an agentive subject, some working-class members of the Legions, with a direct object, open air meetings linked by the transitive verb held.
Similarly, examples (5.3b) and (5.3c) also contain an agentive subject, except for the fact that a metonymic interpretation for the subject is necessary. For both cases, the court represents the decision made by the judges on the case. In (5.3b) earlier, a cognition verb, thought, is used to state how the Court of Appeal, the second highest court in Britain, rationalize the case. Other cognition verbs also found in the data include explain, understand, assume, describe, and maintain.
The majority of the of -constructions exhibit a low degree of human involvement, although we would expect a higher proportion of. Verbs involving physical actions (e.g., come and go) are much less commonly used with the animate of -constructions discussed above. Examples like (5.3c), where exercised, meaning to practice or carry out the law,
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
is used, are rare. This group of verbs is what Biber et al. (1999: 361) refer to as activity verbs and the most common ones found in their academic prose are make, give, take, use, follow, produce, show, and provide (p. 367). This list of verbs are commonly used with inanimate subjects with an occurrence rate of minimally 20 times per million words (p. 378). Combination of this kind was also found in the present study as illustrated in (5.4). As the majority of the of -construction as subject are inanimate entities (384 out of 434, 88%), we expected a large number of instances to follow this pattern.
(5.4) a. However, the statistics are compiled, an examination of local government expenditure shows a steady increase over the years.
(ED5-1314)
b. A synonym occurs whenthe application of the randomising formula to two or more keys gives the same address. (HRK-1897)
c. This use of technology (like Jethro Tull’s seed drill) increased agricultural productivity…(G1G-89)
d. Episodes of acid rain (in the range 3.0-3.6 pH) may have caused major environmental damage to... (GU5-1116)
e. The first decade of the twentieth century witnessed the polarization of these positions under the impact of growing political activism.(G0D-1037)
The inanimate of -constructions as subject illustrated in (5.4a) to (5.4c) represent an event where some kind of action (i.e., an examination, an application, and use) has been carried out on N2. This event, in turn, gives rise to a resultative state in the direct object position. In other words, the use of these action of -construction appears to simplify two propositions into one construction. While (5.4d) establishes a causal relation between episodes of acid rain and major environmental damage that is hedged with the epistemic modal auxiliary may, (5.4e) demonstrates an interesting case where a temporal of -construction is used as subject of the verb witnessed. According to Levin (1993), this sentence represents a kind of alternation which she refers to as
‘oblique subject alternations’ (p.79). She also states that “When the verbs take the oblique subject, the “agent” is no longer expressed” (emphasis original, p.79). An example of time subject alternation, a subclass of oblique subject alternation, taken from Levin (1993: 79) is shown in (5.5).
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
(5.5) a. The world saw the beginning of a new era in 1492.
b. 1492 saw the beginning of a new era.
The alternation involves a change in the subject that results in the reduction of a noun phrase. Compare the two sentences in (5.5) where (5.5b) represents the alternation of (5.5a). The original participant as subject (the world) is replaced by the noun phrase in the prepositional phrase (in 1492). Similarly, we can alternate (5.4e) by inventing a subject and repackaging the temporal of -construction in a prepositional phrase: We witnessed the polarization of these positions under the impact of growing political activism in the first decade of the twentieth century. This alternation appears to be lengthier and places a different emphasis as compared to (5.4e). Oblique subject alternation appears to be quite common with of -constructions as subject. In another subclass, abstract cause subject alternation, Levin presents a list of alternating verbs including assert, confirm, demonstrate, establish, explain, imply, indicate, justify, nullify, obscure, proclaim, predict, prove, reveal, show, and suggest (p.81). This list seems to coincide with a group of verbs called reporting verbs that prevail in academic discourse (e.g., Thompson & Yin, 1991; Thomas & Hawes, 1994). Although the majority of these reporting verbs are used with a that-clause complement (e.g., show that) (Hyland & Tse, 2005; Charles, 2006), some possible candidates are shown in (5.6). The alternatives of the concordances (5.6a), (5.6c), (5.6e), and (5.6g) were reconstructed with an invented subject and a prepositional head (both shown in bold) as shown in (5.6b), (5.6d), (5.6f), and (5.6h).
(5.6) a. Comparative studies of thrombolytic agents have shown differences in efficacy for early coronary artery recanalisation… (HWU-1116) b. Researchers in comparative studies of thrombolytic agents have shown
differences in efficacy for early coronary artery recanalisation…
c. Factor analysis of these images indicated just one general factor underlying image scores for Kinnock, Steel, and Owen…(A62-570) d. They indicated just one general factor underlying image scores for
Kinnock, Steel, and Owen in a factor analysis of these images…
e. That is because the resurgence of development pressures in the more
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
very mixed blessing to the low-income residents which these areas have inherited. (G08-628)
f. That is because we have proved a very mixed blessing to the low-income residents which these areas have inherited on the basis of the resurgence of development pressures in the more central parts of the large cities, particularly London.
g. All forms of office employment reveal a marked concentration in the South East, which has nearly half of the total. (EB2-153)
h. We reveal a marked concentration in the South East, which has nearly half of the total with all forms of office employment.
With the addition of an animate subject, i.e., researchers, we, and they, these alternated sentences become more author-oriented as compared to research-oriented ones found in the original sentences. The use of personal pronouns may also function as cohesive devices to establish a linkage to previous text. For example, the use of they in (5.6d) and we in (5.6h) and that in (5.6f) would imply a previous mention. Overall, a comparison between the alternations suggests that of -constructions play an important role in creating objectivity by demoting the agentivity involved in research action.
Master (1991; 2001) refers to this type of construction as performing a hedging function. Hedging is an important function in academic writing (e.g., Skelton, 1997;
Hyland, 1998). According to Hyland (1998: 3),“hedging is one part of epistemic modality; it indicates an unwillingness to make an explicit and complete commitment to the truth of propositions.” Hyland goes on to explain the importance of hedging in scientific writing: “...scientific argument is part of a rhetoric employed to convince readers of the validity of claims. This search for consensus provides the motivation for the expression of doubt and tentativeness in science for such interpretations can help secure readers’ acceptance” (p.7). To put it simply, academic writers make use of hedging devices to allow smooth negotiation with the readers and of -construction at the subject position seems to facilitate this function.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
5.1.1.2 Object
In terms of of -constructions at the object position (including 1,379 instances of direct object and 428 instances of prepositional object), a total of 1,806 instances are available.
The of -constructions at the object position were further analyzed in terms of their semantic relations and summarized in Table 5.2. This table shows that approximately forty percent of the of -construction as the object belongs to ‘quality’, which is the double of ‘action’ (22.5%) and ‘part-whole’ (21.9%). If we examine the frequency closely for each subcategory, ‘objective’ is the only exception that exceeds each single subcategory under ‘quality’. The frequency of ‘objective’ (20.2%) is higher than
‘attributive’ (14.8%), ‘equative’ (10.8%), ‘topic’ (10.0%), ‘measure’ (9.9%), and
‘containment’ (1.4%). The of -constructions as the object appear to undergo actions of various kinds. The most common ones are perhaps activity verbs including give, take, produce, detect, carry out, exercise, suppress, limit, restrict, allow, share, highlight, bolster, structure, shape, monopolize, surround, improve, and establish. In contrast, reporting verbs (e.g., show, describe, depict) and cognition verbs (e.g., contemplate, know, calculate), which are prevalent when the of -construction occupies the subject position, are much rarer found.
To help elucidate the possible relationships among clause types and semantic relations, a multivariate statistical analysis, hierarchical configural frequency analysis (HCFA) (Gries, 2004), was used to identify significant patterns based on a test of interactions between the following categories: semantic relation subcategory, N1 semantic category, N2 semantic category, and clause type. Statistically significant results applied with a more stringent criteria limiting the effect size measure Q to greater than 0.01 are presented in Table 5.3.
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Table 5.2 Distribution of semantic relations of of -constructions as the object Category Subcategory Count Frequency
Quality attributive 256 14.17%
839 equative 202 11.18%
(46.43%) topic 182 10.07%
measurement 168 9.30%
containment 31 1.72%
collective 44 2.43%
Action objective 408 22.58%
456 (25.24%) subjective 48 2.66%
Part-whole belonging 106 5.87%
415 partitive 99 5.48%
(22.97%) type 86 4.76%
spatio-temporal 70 3.87%
relational 10 0.55 %
Derivative causal 41 2.27%
97 (5.37%) product-source 56 3.10%
Total 1,807 100.00%
‧ 國 立
政 治 大
‧
學tio Na
l C na e h
g n
c h i U n
iv ers i ty
Table 5.3 Significant HCFA results for of -constructions as the object semantic relations N1 category N2 category clause
types
Freq Exp Cont.chisq P.adj.Holm Q
objective act . finite 151 47.5254 225.2899 2.51E-31 0.072
objective act . to V 35 9.8766 63.9073 6.38E-07 0.017
objective act tech/concrete finite 32 7.2671 84.1765 2.54E-07 0.017
collective group . finite 23 1.796 250.3449 7.58E-15 0.014
attributive attribute . finite 78 24.8689 113.5122 1.00E-14 0.037
attributive attribute . Ving 20 3.319 83.8369 7.56E-07 0.011
measurement quantity . finite 27 3.1766 178.6704 2.29E-13 0.016
measurement attribute . finite 52 17.0828 71.3706 1.02E-08 0.024
equative mental . finite 46 21.3595 28.4254 0.00364226 0.017
topic mental . finite 43 19.7576 27.342 0.00597554 0.016
belonging . group finite 21 5.1027 49.5282 0.00018619 0.011
122
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Table 5.3 has nine columns. The first four are the factors taken into consideration for running the HCFA. Each row of these four columns list the the subcategories annotated for each concordance line. For example, the first column is the factor ‘semantic relation’ which has several subcategories including ‘objective’, ‘collective’,
‘attributive’, and ‘measurement’. The last five columns are the number of concordance lines observed (Freq), the expected frequency (Exp), the chi-squared value (Cont.chisq), the adjusted p value (P.adj.Holm), and effect size (Q). The HCFA results were somewhat unanticipated. No significant interaction between semantic relation subcategory and clause type was found. Instead, significant results were those associated with N1 or N2 semantic categories. For example, the first line in the table tells us that the of -construction consists of an ‘act’ N1 (e.g., the acquisition) which forms an ‘objective’ relation with an N2 (e.g., an ability) that follows a finite verb (e.g., results in) to form a finite clause (see (5.7a). This type of ‘objective’ relation with an ‘act’ N1 was also found to be significant with -to infinitive clauses as exemplified in (5.7b). However, the most interesting combination would be shown in the third line of the table where all four factors are involved as found in example (5.7c).
(5.7) a. An aptitude is demonstrated by success in learning which results in the acquisition of an ability. (CLP-1074)
b. If Engles’s book seems in many important ways to follow the lead of Marx’s later works there are also some differences. (ASB-75)
c. In Ehler-Danlos syndrome, Type IV (EDSIV) G to T transition at the splice donor site causes skipping of the preceding exon and production of defective Type III collagen (30). (FTB-407)
Other examples include limit the progression of those parasites, causes a substantial run-off of solids from buildings and roads, apply to the use of nuclear and similar weapons, reflected in the real growth of disposable income, permit release of lysosomal protease. All together, these examples demonstrate an important use of the of -construction in academic writing, that is to express actions carried out on the technical or concrete entities by treating the nominalized action as an object in a reified
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
manner (Schmid, 2000).
Another interesting result is the interaction of ‘attribute’ N1 in an ‘attributive’
relation with N2 with - ing participle clause. This is the only significant interaction involving the - ing participle clause. Instances in (5.8) are some examples.
(5.8) a. A second strategy involved demeaning the importance of money and monetary success. (J14-1389)
b. We can not now implement the whole of that programme and must look for a more concessive approach, capable of accommodating the competing strengths of individualism. (FP2-795)
c. Neitherside should be afraid to instruct more than one doctor when the plaintiff suffers multiple injuries giving rise to a variety of symptoms.
(J6V-133)
As mentioned earlier, the -ing participle clauses are not considered as a typical clausal element in a canonical SOV clause. However, in example (5.8a), the -ing participle clause headed by the verb demeaning serves as the object after the main verb involved.
Example (5.8b) is a case of adjective phrase capable of that precedes the -ing participle clause. Both examples (5.8b) and (5.8c) illustrate a postmodifying function with the participle clause. While the attributive of -construction relates a quality to a second entity or concept, its interaction with the -ing participle clause could reflect the role it plays in providing additional information or characterizing.
5.1.1.3 Complement
Another syntactic position, the complement, is where the most of -constructions were found in the academic corpus data with a total of 41 percent. There are mainly three types of complement divided in this study, namely, subject complement (5.9a), object complement (5.9b), and complement of preposition (5.9c).
(5.9) a. This project is an analysis of the survey material already available on the subject of adoption of, and resistance to, Information Technology in the workplace. (HJ1-23247)
b. In the United Kingdom many midwives, obstetricians, and mothers would consider this an unacceptably high level of intervention in an otherwise normal physiological process. (FT5-930)
‧ 國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
c. A further group of medium-sized districts sought less extensive changes
c. A further group of medium-sized districts sought less extensive changes