研究論文中報導動詞之語意與言談功能

行政院國家科學委員會補助專題研究計畫成果報告

※※※※※※※※※※※※※※※※※※※※※※※※※

研究論文中報導動詞之語意與言談功能

Reporting Verbs : Semantics and Discourse Functions

in Research Papers

※※※※※※※※※※※※※※※※※※※※※※※※※

計畫類別：v 個別型計畫

□整合型計畫

計畫編號：NSC89－2411－H－009－022

執行期間：89 年 8 月 1 日至 90 年 7 月 31 日

計畫主持人：郭志華

共同主持人：

計畫參與人員：

執行單位：國立交通大學語言教學與研究中心

中華民國九十年十月三十一日

行政院國家科學委員會專題研究計畫成果報告

研究論文中報導動詞之語意與言談功能

Repor ting Ver bs :

Semantics and Discour se Functions in Resear ch Paper s

計畫編號：NSC 89-2411-H-009-022

執行期限：89 年 8 月 1 日至 90 年 7 月 31 日

主持人：郭志華 國立交通大學語言教學與研究中心

一、 中文摘要 本計畫探討期刊論文中之引述所用之 報導動詞。研究採用質與量的分析，首先 找 出 四 種 期 刊 48 篇 論 文 中 的 引 言 (Introduction) 及 討 論 與 結 論 (Discussions and Conclusions) 二章節中共 877 個引述和 174 個報導動詞，並依史威爾氏(Swales)之 建議，以引述的研究或研究者是否為論文 內文一部分以及引述中是否使用報導動詞 兩個原則來將引述分為四類 ─ 整合式報 導 (integral-reporting) 、 整 合 式 非 報 導 (integral-nonreporting) 、 非 整 合 式 報 導 (nonintegral-reporting) 、 非 整 合 式 非 報 導 (nonintegral-nonreporting) ─ 以 探 討 不同 引述形式在論文中之常用程度與功能。其 次，本研究比較了論文的引言和討論與結 論二章節中出現的引述與報導動詞使用上 之異同。最後，再比較社會科學與科技領 域期刊論文中出現的引述與報導動詞使用 上之異同。 研究結果顯示，論文中含報導動詞之 引述多於不含報導動詞之引述，同時整合 式引述比非整合式引述較常使用報導動 詞。引言所含的引述與報導動詞均多於討 論與結論。社會科學領域論文所含的引述 與報導動詞均多於科技領域論文，且二者 使用的報導動詞亦有相當差異，反映了學 術領域性質上之不同。 另一有趣的分析結果是科技領域論文 中出現自我引述(self-citation)之比率遠高 於社會科學領域之論文，而在後者中出現 含直接引語之引述(direct-quotation citation) 的比率則遠大於前者。研究報告中並討論 了上述結果之可能原因及其與論文之溝通 功能之關係。 由於非母語人士對於撰寫期刊論文的 引述常覺困難，本計畫之研究結果提供了 各類型引述及報導動詞之用法，具有論文 教學上之實用價值。 關鍵詞：引述、報導動詞、整合式報導、 整合式非報導、非整合式報導、非整合式 非報導、自我引述、含直接引語之引述 Abstr act

This project investigated the use of re-porting verbs in citations in research papers. Both qualitative and quantitative analyses were conducted. In total, 877 citations and 174 reporting verbs were identified in the Introduction and the Discussions and Con-clusions sections of 48 research papers in four major journals. Based on Swales’ (1990) proposal, the citations were grouped into four categories – integral-reporting, inte-gral-nonreporting, nonintegral-reporting, and nonintegral-nonreporting – and compared for their occurrences and functions in research papers. Furthermore, reporting verbs and ci-tations in the Introduction section were com-pared with those in the Discussions and Con-clusions section. Finally, a comparison was also made of the reporting verbs and citations in the research papers in social sciences with those in science and technology.

Results from the analyses show that there aare more reporting citations than non-reporting citations. Also, non-reporting verbs oc-cur more frequently in integral citations than in nonintegral citations. It is also found that the Introduction section contains more

re-porting verbs and citations than the Discus-sions and ConcluDiscus-sions section. On the other hand, the research papers in social sciences contain more reporting verbs and citations than those in science and technology. In ad-dition, a large percentage of the reporting verbs occurring in the former are different from those occurring in the latter in terms of their evaluative nature, reflecting disciplinary difference.

Another interesting result from the analyses is that research papers in science and technology use self-citations far more frequently than those in social sciences, while the latter use direct-quotation citations far more frequently than the former. Possible interpretations of the results and their rela-tionships with the communicative purposes of the research paper are discussed in the re-port.

Nonnative writers of research papers often have problems writing citations. The results of this project should provide valuable information about the use of reporting verbs as well as citations in different sections of the research paper and in different disciplines.

Keywor ds: citation, reporting verb,

gral-reporting, inte-gral-nonreporting, gral-reporting, noninte-gral-nonreporting, self-citation, direct-quotation citation 二、緣由與目的 (Intr oduction)

Reporting verbs refer to verbs used in reporting, or citing, others’ research in aca-demic research papers. Citations may be pre-sented in various forms and perform various discourse functions in a research article (Swales 1990). There has been much interest in this research area. A number of studies have tried to establish a theory of citations (Gilbert 1977; Small 1978; Cronin 1981; Leydesdorff 1987; Leydesdorff & Amster-damska 1990), while others have focused on the classification of citations in terms of their forms or functions (Moravcsik & Murugesan 1975 ; Chubin & Moitra 1975; Peritz 1983; Swales 1986; Dubois 1988).

The results from the above studies sug-gest the strategic nature of citations at the discourse level. More recently, investigations have been made into how various reporting structures in citations are signaled or repre-sented through the choice of reporting verbs as well as their tense, voice, or syntactic structures (Oster 1981; Lackstrom, Selinker & Trimble 1972; Malcolm 1987; Shaw 1992; Tarone et al. 1981; Thompson & Ye 1991; Thomas & Hawes 1994). For example, as Swales (1990:154) indicated, “Whatever the reasons, the tense choice may indicate some-thing of the author’s stance towards the cited work, and … .” The progression from past to present perfect to present is shown as a kind of strategic tense choice suggesting increas-ing proximity (Een 1982; Malcolm 1987; Lackstrom, Selinker & Trimble 1972; Swales 1981).

Then, a reporting verb itself in a report-ing statement could carry the writer’s evalua-tion or degree of commitment towards the cited study (Thompson and Ye 1991; Swales 1990). The semantics of reporting verbs in relation to their rhetorical functions could be identified and categorized to provide options and patterns of choices (Thomas & Hawes 1994).

Shaw (1992), on the other hand, argues that the correlation between tense, voice, and sentence function in reporting verbs could be explained in terms of thematization; in other words, the choice is based on the rhetorical function of the sentence to set the theme or meet the cohesion requirements in the re-search paper.

Bazerman (1984), examining the style and form of Physical Review chronologically,

found that there was a decrease in reporting verbs and increase in active verbs. He sug-gests that the finding or theory has played a more important role in the experimental re-port as it has been given the grammatical subject position more frequently.

Following Shaw’s and Bazerman’s findings, we suspected that genre or dis-course requirements in the research paper determine largely the forms of citations, which in turn constrain the choice of report-ing verbs. This study, therefore, examined

empirically the use of reporting verbs in two forms of citations: integral and non-integral citations (Swales 1990:148).

In addition, it has been found that cita-tions are no longer concentrated in the Intro-duction section but are distributed throughout the research paper (Swales 1990; Chang & Chang 1999). It is believed that citations and also reporting verbs in different sections of the research article perform different dis-course functions. The choice and use of re-porting verbs in the Introduction section and the Discussions and Conclusions section, therefore, were compared.

Finally, in addition to the use of differ-ent citation systems (Swales 1990), differdiffer-ent patterns of thematic development and differ-ent natures of knowledge presdiffer-entation in dif-ferent disciplines could lead to the use of different forms of citations and reporting verbs. We, therefore, further compared the reporting verbs in research papers in social sciences (SS) and those in science and tech-nology (ST).

Forty-eight research articles were ran-domly selected from four major journals, two from SS (Applied Linguistics, English for Specific Purposes) and the other two from ST

(IEEE Transactions on Computers, IEEE Journal of Quantum Electronics). Review

papers were excluded as they have a different nature in terms of citations. Each citation in the Introduction section (IN) and the Discus-sions and ConcluDiscus-sions section (DC) of these articles was identified as integral (I) or non-integral (NI). Then, each was further identified as reporting (R) or non-reporting (NR). The identification of the four catego-ries of citations – IR, NIR, INR, and NINR – was based on Swales’ definitions (Swales 1990: 148-150).

Frequency analysis was done for each of the four categories of citations, and for each of the reporting verbs used in the two catego-ries (IR and NIR) where reporting structures were identified. Self-citations and citations with direct quotations were also counted re-spectively.

A detailed analysis of the various types of sentence structures that occur in integral and non-integral citations was made; in

par-ticular, citations with the researcher’s name as the grammatical subject and citations not using the researcher’s name as the gram-matical subject were compared.

Then, the four different categories of ci-tations as well as the reporting verbs occur-ring in the Introduction section (IN) and the Discussions and Conclusions section (DC) were compared.

The last step was to compare the cita-tions and reporting verbs in the research pa-pers in social sciences with those in science and technology.

三、結果與討論 (Results and Discussions)

Both quantitative and qualitative analy-ses were made of the citations and reporting verbs in the sample research papers. In the following, however, only major findings from the analyses are presented as a result of the limitation of space in this report. It is also impossible for us to present detailed data in tables and figures.

Integral(I)/Non-integral (NI) and

Reporting(R)/Non-reporting (NR) Citations

In total, 877 citations were identified. No significant difference between the number of I (444, 50.6%) and NI (433, 49.4%) was found. However, more reporting structures (619, 70.6%) were found than non-reporting structures (258, 29.4%) in citations. Fur-thermore, integral citations used more re-porting structures (i.e., IR, 378) than non-integral citations (i.e., NIR, 241). Among the four categories of citations, IR (378) occurred most frequently, followed by NIR (241), NINR (192), and INR (66) oc-curred least frequently.

There were totally 111 self-citations (12.7% of all citations), and 116 di-rect-quotation citations (13.2% of all cita-tions).

Introduction (IN) vs. Discussions and Con-clusions (DC)

Much more citations were found in IN (691) than in DC (186). However, a further

analysis revealed that there were more inte-gral citations (IR + INR, 360) than non-integral citations (NIR + NINR, 331) in IN, while there were more non-integral cita-tions (102) than integral citacita-tions (84) in DC.

In both IN and DC, there were more re-porting citations (510 and 109 respectively) than non-reporting citations (181 and 77 re-spectively).

In addition, it was found that there were more self-citations in IN (77) than DC (34), and also more direct-quotation citations in IN (93) than DC (23).

Social sciences (SS) vs. science and technol-ogy (ST)

A comparison of the numbers of cita-tions of the various categories between SS and ST clarified and modified our interpreta-tion of the data presented previously, since we could see disciplinary difference plays a role in citation practice.

First, much more citations occurred in SS (639) than ST (238). This was true in both IN and DC.

In terms of citation types, integral cita-tion was preferred to non-integral citacita-tion in SS (398:241 or 62.3%:37.7%), but there were more non-integral citations than integral cita-tions in ST( 192:46 or 80.7%:19.3%). Al-though it is possible that a numerical citation system, which is usually adopted by journals in ST, does not easily permit integral cita-tions (Swales 1990:151), we suspect that dif-ferent styles of information presentation and argumentation in SS and ST may also lead to the use of different types of citations, as in-dicated in the previous section. A further analysis of the reporting structures and the reporting verbs used in SS and ST shed more light on this point, which will be discussed later.

With respect to reporting or non-reporting citations, both SS and ST used more reporting (447 and 172) than non-reporting citations (192 and 66).

It should also be noted that self-citation occurred much more frequently in ST (69, 29%) than in SS (42, 6.6%). Furthermore, a higher percentage of self-citations occurred

in DC than in IN, as indicated earlier. There may be two possible reasons for such results. First, as a result of the continuing nature of their research, ST researchers often need to discuss the findings and results of the present research in relation to their own previous studies. Another interpretation is that citing one’s own research is “an important means of demonstrating one’s disciplinary credentials and credibility.” (Hyland 2001:214) In con-trast, self-citing did not, obviously, occur of-ten in SS. It is possible that issues in SS of-tend to be relatively diverse and there is compara-tively little opportunity for self-citation (Becher 1989; Hyland 2001). In addition, it could be that researchers in SS do not prefer blowing their own trumpets. Rather, citation is a practice mainly for the purpose of sup-porting one’s perspectives and propositions, In SS, this would be better achieved by citing other researchers, particularly the more well-established scholars in the field.

On the other hand, research papers in SS (112, 17.5%) used direct-quotation citations much more frequently than those in ST (4, 1.7%). This also suggests the underlying dif-ference of knowledge presentation and style of argumentation between ST and SS.

Reporting verbs

In total, 174 reporting verbs were found in the IN and DC of all sample research pa-pers. There were 162 reporting verbs in IN and 47 in DC. Thirty-five reporting verbs occurred in both IN and DC. In terms of fre-quency, the top 10 reporting verbs were: find

(57), show (26), see (25), suggest (23), report

(16), use (15), indicate (15), point (out) (13), note (12), demonstrate (11), describe (11). A

further comparison of the respective lists of the top 10 reporting verbs in IN and DC re-vealed that use, point (out), demonstrate, de-scribe, and propose did not appear in the

lat-ter while discuss, argue, and confirm did not

appear in the former.

These results suggest: firstly, although reporting verbs may occur in other sections of the research paper, they occur much more frequently in IN since reporting others’ re-search, or literature review, is a major move

in IN; secondly, the more frequently used re-porting verbs are those rere-porting the cited study in a general, non-evaluating, and non-committing way. Finally, some argument verbs or cognition verbs (Thomas & Hawes 1994) may often occur in DC as a result of the discourse functions of DC in the research paper.

With respect to reporting verbs in SS and ST, 174 reporting verbs were found in total; however, there were 139 in SS, and 77 in ST, since 42 reporting verbs (24.1%) oc-curred in both SS and ST, 35 ococ-curred in ST only, and 97 occurred in SS only. Despite the small sample size, it is of interest to find that reporting verbs referring to scientific research activities such as calculate, configure, fabri-cate, feature, incorporate, infer, implement, perform, and solve occur in ST only, while a

lot of reporting verbs occurring in SS only, as can be noted, seem to be argumentative, cog-nitive, or evaluative, such as agree, attribute, argue, believe, contend, caution, imply, in-dict, influence, lament, comment, maintain, motivate, overlook, posit, promise, speculate,

and stress, to name a few. The results, again,

may be linked to disciplinary differences, which are marked in genre conventions.

The occurrences of reporting verbs in integral and non-integral citations were also compared. It was found, as may be expected, that there were more reporting verbs in inte-gral (131) than in non-inteinte-gral citations (99).

Syntactic forms of citations

The syntactic structures occurring in both integral and non-integral citations were analyzed; in particular, the number of two categories of structures were counted: those in which a researcher’s name (or a citation number which represents a cited study) serves as the grammatical subject of the sen-tence (S), and those in which the grammatical subject is not a researcher’s name or a cita-tion number (NS). It was found that, as ex-pected, S did not occur in non-integral cita-tions, while in integral citacita-tions, S occurred more frequently than NS. However, it was also found that although S occurred more of-ten than NS in IN, NS occurred more ofof-ten

than S in DC. A possible reason is that an important communicative function of DC is to interpret, discuss, or summarize the major results from research; hence, the focus would be placed on the related information content from the cited study rather than its researcher. NS is more appropriate than S for such in-formation-oriented citations. In other words, citations in a certain section of the research paper may be realized more often through certain syntactic structures than other struc-tures as a result of the specific communica-tive purposes of the section in the research paper.

In terms of structural variety, the fol-lowing pattern is typical in S: Kuo and Fuchs (3) showed… . (or (3) proposed … .) But NS

shows more structural variety. A number of patterns often occur in:

(a) NS in integral citations

*According to Clark and Hecht (1983), … .

*… , as has been suggested by Levelt, Sinclair, and Jarrella (1978), … .

*The model in (10) provides … .

*These studies follow on earlier studies such as Long (1981, 1983), … .

*This is confirmed by … developed by Tromborg et al. (4).

*In (10), … .

(b) NS in non-integral citations

*Researchers have found … (Hatch 1974; Cancino, Rosansky, and Schumann 1978; Felix 1980a).

*It was found that … (see Scollon & Scollon 1981 ).

*A number of studies have compared … (3-7).

*There has been increasing interest in … (1-7).

四、計畫成果自評 (Self-evaluation)

This project investigates reporting verbs in the citations of the research paper. The re-sults show that the use of reporting verbs is associated with the syntactic form of a cita-tions, communicative funccita-tions, and disci-plinary conventions. There are significant differences between SS and ST in terms of

the number and types of citations as well as specific reporting verbs. There is also a sig-nificant difference between IN and DC in the number of citations and reporting verbs.

The analyses mainly rely on human reading, thus taking much time. The sample size, therefore, is not very large and gener-alizations from the results must be cautious.

On the other hand, this study is a further step from previous studies, which have fo-cused on rhetorical or semantic classification of citations and reporting verbs. We examine reporting verbs in two discourse types of ci-tations. We also compare reporting verbs and citations in two major sections, which have different discourse functions in the research paper. Two disciplines of different nature are compared as well. The results, therefore, provide specific information for a better un-derstanding of the use of reporting verbs and citations in different sections of the research paper as well as different disciplines.

五、參考文獻 (References)

Bazerman, C. (1984). Modern Evolution of the Experimental Report in Physics: Spectroscopic Articles in Physical Re-view, 1893-1980, Social Studies in Sci-ence, 14, 163-196.

Becher, T. (1989). Academic Tribes and Ter-ritories: Intellectual Inquiry and the Cultures of Disciplines. Milton Keynes:

SRHE/OUP.

Chang, Y. L. and Chang, T. S. (1999). Cita-tion Analysis in Journal Articles. In The Proceedings of the 1999 English for Specific Purposes Conference, Ming

Chuan University, Taoyuan.

Chubin, D. E. and Moitra, S. D. (1975). Content Analysis of References ： Adjunct or Alternative to Citation Counting? Social Studies of Science, 5,

423-441.

Cronin, B. (1981). The Needs for a Theory of Citing, Journal of Documentation, 37, 1,

16-24.

Dubois, B. L. (1988). Citation in Biomedical Journal Articles, English for Special Purposes, 7, 181-193.

Gilbert, G. N. (1977). Referencing as

Persua-sion, Social Studies of Science, 7,

113-132.

Hyland, K. (2001). Humble Servants of the Discipline? Self-mention in Research Articles, English for Specific Purposes,

20, 3, 207-226.

Lackstrom, J., Selinker, L., and Trimble, L. (1972). Grammar and Technical Eng-lish, English Teaching Forum, X, 5,

3-14.

Leydesdorff, L.(1987). Towards a Theory of Citation, Scientometrics, 12, 287-291.

Leydesdorff, L. and Amsterdamska, O. (1990). Dimensions of Citation Analy-sis, Science Technology and Human Values, 15, 305-335.

Moravcsik, M. J. and Murugesan, P. (1975). Some Results on Function and Quality of Citations, Social Studies of Science,

5, 86-92.

Oster, S. (1981). The Use of Tenses in Re-porting Past Literature. In Selinker, Tarone, & Hanzeli (Eds.), English for academic and Technical Purposes: Studies in Honour of Louis Trimble.

Newbury House.

Peritz, B. C. (1983). A Classification of Cita-tion Roles for the Social Sciences and Related Fields, Scientometrics, 5,

303-312.

Shaw, P. (1992). Reasons for the Correlation of Voice, Tense and Sentence Function in Reporting Verbs, Applied Linguistics,

13, 302-319.

Small, H. G.(1978). Cited Documents as Concept Symbols, Social Studies of Science, 8, 327-340.

Swales, J. (1981). Aspects of Article Intro-duction. Birmingham, U. K. : The

Uni-versity of Aston, Language Studies Unit.

Swales, J. (1986). Citation Analysis and Discourse Analysis, Applied Linguistics,

7, 1, 39-56.

Swales, J. M. (1990). Genre Analysis: Eng-lish in Academic and Research Settings.

Cambridge: Cambridge University Press.

Thomas, S. and Hawes, T. (1994). Reporting Verbs in Medical Journal Articles,

129-148.

Thompson, G. & Ye, Y. (1991). Evaluation in the Reporting Verbs Used in Aca-demic Papers, Applied Linguistics, 12,