The findings of the analysis above indicate that although there are shared writing conventions across different disciplines in the academic discourse community, unique practice in the choice of active/passive voice may characterize hard and soft science.
For example, passive voice is generally believed to be a technique for researchers to reduce human involvement in research articles. The choice between active and passive
voice in hard and soft science as shown in the quantitative analysis reveals different patterns. We, therefore, conducted qualitative analysis in order to learn (1) possible explanations for the results of the quantitative analysis; (2) whether factors other than objectivity contribute to voice choice in research articles. We examined the contexts where passives or actives occur and compare them, especially in relation to the theme and the information system of text, as mentioned earlier.
Voice Use in JMS
In JMS, the extensive use of passive voice in the Methods section (58.98%) reflects the general concept that objectivity is paramount in describing the research methodology and the experimental process. This can be verified from three findings in our examination of the nature of agents in passive voice structure. Firstly, only 3 occurrences (2.36%) of animate agent were found in all by-agents. Secondly, there are a large number of occurrences of agentless passives (73.15%); this suggests writers might try to focus on the research per se. Finally, when by-agents are used, they are inanimate agents shown as “methods” and “equipment/materials,” indicating the means employed in research to detail the experimental procedure in order to increase the validity of research.
Voice Use in ESP
In contrast to JMS, ESP has more active than passive constructions (68.93% to 31.07%). Further examining the content of the sample texts, we found that there may be three possible reasons for this: firstly, unlike hard science in which inanimate materials, equipment, model, etc. usually play essential roles, soft science often has people as subjects, who are able to “actively” perform actions. For example:
[4.15] ...the native-speaking US students at MSU and the Catalans at UHC share characteristics that separate them from the US immigrants.... The first two groups present the traditional profiles of public university undergraduates in their countries.... If members of either group were to study at the others’
institution, they would be visa students. On the other hand, the immigrant students at UCC receive a lesser degree, the Associates, upon graduation....
They belong to a minority ethnic group; they come from lower social-class backgrounds...; and they have inferior secondly education in their home countries and in the USA. (Text 11b)
[4.16] In this research the sampled students are in their early twenties and come from the degree of ‘Business’, ‘Chemical Engineering’, ‘English Philology’
and ‘Telecommunications’. They are representative of Spanish university undergraduates studying many subjects, with lectures and exams for each....On average, they have studied English for at least 5 years before entering university. (Text 1b)
In [4.15] and [4.16], subjects under investigation take the position of the sentence subject and, at the same time, active voice is applied. This suggests the extensive use of active voice is mainly due to the fact that animate subjects are used more often in the descriptions of research method in soft science.
The second explanation is that writers in the field of social science usually describe how they conduct their investigation as if they were addressing the readers.
Followings are passages from our data:
[4.17] We chose, in our present study, to observe and videotape students’ beside performance in wad teaching, as we felt that this approach would offer the most direct, reliable, and rapid access to the kinds of problems with which we were concerned. For our purposes, we defined students’ needs or problems as
“elements of competent academic.... Specifically, we focused on those language errors that caused communication problems. ..., we consulted clinical training textbooks for information on how to interview patients,....
We recorded six ward-teaching sessions over a period of three months....(Text
24b)
[4.18] ..., I reviewed several inventories of questions and found Christensen’s typology the most appropriate.... I supplemented it with several categories of particular relevance to second language teaching situations. Table 1 shows the typology I employed, along with examples from my database.... In the transcripts, I then identified all of the case leaders’ questions and classified them using the typology. On the code sheet, I also indicated whether each question was open-ended or closed, according to the definitions shown in Table 2. (Text 16b)
In [4.17], writers adopt We-structure to justify and explain their research method while in [4.18], sentences are even initiated with I to describe the research process. The use of first-person pronouns, therefore, contributes to the choice of active voice.
Finally, in addition to the involvement of researchers and test-takers in the Methods section, informants’ or experts’ perspectives are often brought up in the field of social science in order to present the theoretical background of the present study or to show the methodological significance of the present study by indicating support from these informants or experts. For example:
[4.19] Bhatia (1993) has the most complete discussion of letters of application as a genre. He found a six-part structure in his study of letters of application for academic positions at universities.... Further, Bhatia noted that inadequate textualizations in the application letter could lead to failing to get short-listed for a job. ...Bhatia pointed out that a significant difference between
South-Asian application letters.... (Text 21b)
In [4.19], in order to justify their own research, the writers put Bhatia, the well-known researchers in genre study, in the subject position to introduce Bhatia’s claims in four sentences. Active voice is naturally applied.
As indicated above, we found that the extensive use of active voice in the research articles of social science can be attributed to the animate nature of sentence
subjects, and their practical function to facilitate communication with readers and to render support to the authors’ propositions.
Potential Factors of Voice Choice
The qualitative analysis also reveals that objectivity is not the only reason why researchers choose to encode sentences in passive voice rather than active voice.
Theme and information systems seem to be involved. For example,
[4.20] High purity powders of BaCO3 (>99.5%), ND2O3 (>99.5%), ZnO (>99.95%) and M2O5 (M = Nb, Ta) (>99.9%) were weighted to Ba5NdZnM9O30 (M = Nb, Ta) stochiometry. The compounds with M as NB and Ta will be referred to as BNZN and BNZT, respectively in the remaining part of the text. The powders were ground in agate mortar to obtain homogeneous mixtures and calcined in 1350 and 1450 ℃ for 48h for BNZN and BNZT, respectively.
The calcined powders were reground into very fine powders and palletized into disks of 11 mm diameter and about 2-4 mm thickness using a cold isostatic press with a pressure of 200 MPa. (Text 2a)
[4.21] The data for this study comprise the Results, Discussion and other following sections of 20 RAs that report empirical research. These articles form part of a corpus of 40 RAs (the remaining RAs did not report empirical studies). The 40 articles had been selected on a stratified random basis (10 RAs per journal were chosen, following reference to a table of random numbers) from four established journals of applied linguistics. (Text 5b)
In examples [4.20] and [4.21], note that “the powders” and “articles” are used as grammatical subjects in a string of sentences since they are the theme foci in each paragraph, and therefore fronted to the subject position. The use of voice, as shown in each example, is influenced by the theme system.
In some other cases, the information system, that is, the given-to-new principle seems to govern voice choice. For example,
[4.22] For the electric test, raw samples were conformed into disk pellets and then fired at 1000 ℃ for 48h, 1150 ℃ fro 100h, 1300 ℃ for 24h and 1400 ℃for 10h. The palletized samples were electroded onto both surfaces with silver high-conductivity paste. These contacts were fired at 300 ℃ for 5 h. (Text 3a)
[4.23] The survey was distributed in 1998 to employers of Monash graduates with business degrees (Bachelor of Commerce and Bachelor of Business). The relevant employer companies were derived from university graduate destination survey data, and from the list of companies that attend Monash for careers fairs. Out of a pool of 96 relevant companies, 24 completed the survey. (Text 20b)
Example [4.22] above introduced how the samples were handled step by step: the first sentence described how the raw samples (given information) were conformed into
“disk pellets” (new information). Then, in the next sentence, these “palletized samples” appeared as given information to introduce the next step (“The palletized samples were electroded…”). Likewise, the first sentence in example [4.23] presents
“employers” as new information. Then, in the second sentence this new information, being transformed into a noun phrase (“employer companies”), becomes given information and the grammatical subject of the sentence to elicit the companies qualified in the experiment (new information), and “24” (of these companies) becomes the given information in the next sentence.
Since theme and information systems play important roles in the determination of subjects and voice choice, two questions worth discussing. Firstly, are there any principles governing the application of either theme or information system when writers choose subjects of sentences? Secondly, if there are principles, can they be applied to research articles in both hard as well as soft science? As the use of the same subject in two or more successive sentences can develop subject continuity, writers
often use unmarked theme, that is, combining theme with the grammatical subject, to create a focus in a paragraph or a number of paragraphs. In our analysis of JMS, we found that theme system is usually applied in presenting the method, describing the materials or equipment, like “powders” in example [4.20]. In ESP, the paragraphs describing the characteristics of subjects or texts for analysis are also governed by theme system. On the other hand, the given-to-new principle of information system usually governs the string of sentences depicting the experimental procedures in both JMS and ESP, like examples [4.22] and [4.23]. Thus, we may generalize that theme system is applied in order to focus on one particular theme while the use of information system is often aimed to describe the procedure of investigation.
Functions of By-Agents
With respect to by-agents, we may interpret the results from a functional perspective. Firstly, by-agents are normally considered as the contributors of actions.
Nevertheless, in the study, we noted that by-agents used in research articles have other functions. By-agents can act as a way to explain experimental variables, methods and equipment like “by the type of science,” “by adjusting the heating power” and “by a He-Ne laser.” Secondly, the functions of by-agents in research articles of different natures across disciplines may be different. In the study, most expressed agents in JMS are inanimate agents (97.64%) which serve to illustrate the method and equipment used in the study. In ESP, however, over half of them (54.17%) are animate agents. They are often aimed to show subjects’ and researchers’ performance during the investigation or justify research results by mentioning others’ research outcomes. The functions and patterns of by-agents are summarized in Table 4.9.
Table 4.9 Functions and patterns of by-agents
Functions of by-agents Patterns of by-agents
Animate agents
Subjects Researchers
Writers reviewing literature Others
by + noun (s)/noun phrase (s)
Materials Equipment Factors
by + noun (s)/noun phrase (s)
Inanimate agents
Methods
by + noun (s)/noun phrase (s) by + Ving
by means of + noun (s)/noun phrase (s)
Besides, by in the pattern “by + Ving” is often omitted. For example,
[4.24] The key and termination choices were identified and interpreted within the systems given in Brazil’s model and transcribed using those conventions.
(Text 4b)
[4.25] The density of sintered pellets were measured with the Archimedes method and calculated using the following equation ρ= W air / (W air － W aqu).
(Text 6a)
According to our analysis, the omission of by in the “by + Ving” is extensively used in the genre of journal articles although other types of writing seem not to apply this by-omission pattern.