English for Specific Purposes 23 (2004) 4566 www.elsevier.

com/locate/esp

The use of we in university lectures: reference and function

Abstract The use of rst and second person pronouns is an important indicator of how audiences are conceptualized by speakers and writers in academic discourse. Several grammarians and linguists have studied more or less specically this part of the English language. Two of the most widely discussed aspects of pronouns are their reference and their discourse function. In this paper I present research on the use of the pronoun we in academic speech. I selected this pronoun since a previous study had signalled it as the most frequent in this type of language. That study was conducted by Rounds in 19851987 using a corpus from the University of Michigan. The present research was carried out using the MICASE (Michigan Corpus of Spoken Academic English), also from the University of Michigan. This paper reveals important results about the use of we in English academic oral discourse, especially related to reference and discourse function. Some of these results seem to contradict those obtained by Rounds, whereas others support her research and complement it. # 2003 The American University. Published by Elsevier Ltd. All rights reserved.

1. Introduction The use of rst and second person pronouns is an important indicator of how audiences are conceptualized by speakers and writers in academic discourse, even in modern academic English with its very limited repertoire of I, we and you. In lectures and comparable speech events, they are perhaps specially marked as rhetorical indicators because, at least in English, levels of attempted rapport and degrees of personal involvement can be traced by, say, the choice of I versus we or we versus you.

* Corresponding author. Tel.: +34-964-229287/729613. E-mail address: fortanet@l.uji.es (I. Fortanet). 0889-4906/03/$22.00 # 2003 The American University. Published by Elsevier Ltd. All rights reserved. doi:10.1016/S0889-4906(03)00018-8

46

I. Fortanet / English for Specic Purposes 23 (2004) 4566

One of the most widely discussed aspects of pronouns, including I and you, is their reference. Halliday (1994, pp. 309312) considers pronouns as one of the devices used to create cohesion in English; pronouns, together with demonstratives and comparatives, give reference points for the reader or hearer to understand a speech event. Those reference points can be people or objects in the environment, or previous parts of the text. Among the pronouns that refer to elements of the environment, it is assumed that I and we represent the speaker or writer, and you the addressee (hearer or reader). However, the referent of these pronouns is not always so clear. As Biber, Johansson, Leech, Conrad, and Finnegan (1999) point out, the meaning of the rst person plural is often vague (p.329) or the second person you is similar to we in being used with dierent intended referents. This vagueness has led these authors to state that it is usually left to the addressee to infer who is included in the reference (p.330). In the negotiation of meaning that is always present between the person issuing a message and the person receiving the message one of the key elements is the reference of the personal pronouns. Among the personal pronouns, the ones that are especially important for communication are the rst and second person pronouns for the implications they have for both participants in the speech event. Previous research on pronouns has also focused on discourse function. Some researchers have dealt with this question from the point of view of politeness. For Brown and Levinson (1987, p. 127), for example, the use of we including both speaker and hearer is identied with positive politeness. On the other hand, one employed as a substitute for you and I may have on most occasions the eect of distancing, causing negative politeness. Along the same lines, relating the choice of pronouns to the intended creation of distance between the speaker and the receiver of the message, Kamio (2001) highlights the gradation of closeness from we (highest closeness) through you to they, which he considers as psychologically very distant (pp. 11201121) both from the speakers and the hearers territory. However, he goes on to recognize that even with an individual pronoun, such as we, the reference is not xed and can change from one example to the next, varying also the degree of closeness involved. Kamio may be right in pointing out the gradation of closeness provoked by the use of pronouns but only the context can give enough information to tell if the choice of one or other pronoun is related to a greater or a lesser closeness; and sometimes, even a hearer who is acquainted with most of the elements of the linguistic and extra-linguistic context has to ask the speaker explicitly about their intention in order to understand the message. Some other researchers have found more specic functions of the use of pronouns. Hyland (2001), for example, attributes the use of self-mention by research articles authors to their intention to be closely associated with their work or to mediate in the relationship between their arguments and their discourse communities (p.223). He points out the growing preference of the use of I over we specically in the hard science disciplines. Tang and John (1999) also focus on the referent of the rst person, but using the perspective of postgraduate students academic writing. Following Cherry (1988),

I. Fortanet / English for Specic Purposes 23 (2004) 4566

47

they distinguish three dierent roles adopted by a writer producing a piece of writing, which can be the referent of I in dierent contexts (p.26):  Societal role: identities inherent to a person (mother, father, son, daughter, American, etc.)  Discourse role: identities that a person acquires by participating in a particular discourse community (lawyer or client in a legal discourse community, teacher or student in an educational environment, etc.)  Genre role: specic to a particular genre within the discourse community (writer as a guide through the essay or the architect of the essay). Although these studies have begun to illuminate the functions of pronouns in written texts, none focuses on spoken English. Raising awareness of pronouns in oral discourse may be especially relevant as academic speech is gaining importance due to the internationalization of both graduate and undergraduate studies. Reective of this new interest is the creation of a specic corpus of academic spoken English, the MICASE (Michigan Corpus of Academic Spoken English), which is also helping to encourage its research. The study of pronouns in academic speech is also relevant given the importance of lectures in the lives of non-native English speakers at English-medium universities. As stated by Flowerdew (1994, p.1), referring to universities in English speaking countries, from among the many instructional media at the disposal of teachers . . . the lecture remains the central instructional activity. Flowerdew also points out that researchers in second language learning consider comprehension as a two-stage process, the rst stage consisting of purely linguistic processing and the second of application of the results of this linguistic processing to background knowledge and context (1994, p.9). Therefore, contrary to what students often think, on many occasions it is not the linguistic skills of the hearers that are a hindrance to their understanding, but the application of language to real situations, especially dicult in the case of pronouns, due to their dependence on context. Even for inexperienced or novice lecturers or those who are non-native speakers of English it is important to be able to apply their language skills to the context with eec nez, 1998). tiveness and condence (Gime One of the primary sources of information about rst and second pronoun use in academic speech is Rounds dissertation (1985) and her two subsequent articles (1987a, 1987b) both from the point of view of discourse analysis and of teacher training. These two articles have been quite widely cited in journals that are included in the ISI databasewith 16 citations of her TESOL Quarterly article, and 5 of her English for Specic Purposes article. Rounds (1985, 1987a, 1987b) analysed the use of pronouns in academic lectures. Specically, her corpus consisted of ve 50-min videotapes of native and nonnative English-speaking teaching assistants of mathematics delivered at the University of Michigan, with a total of 26,734 words. For our purposes, her most striking ndings relate to the high frequency of we in instructor speech and the multiple functions of this pronoun in academic talk. She found that, in her corpus,

48

I. Fortanet / English for Specic Purposes 23 (2004) 4566

we occurred up to three times more frequently than I or you, and that the most successful teaching assistants were the ones that made a greater use of we (1987b, p. 648).1 Rounds (1985, p. 4655) also presented semantic mappings for the pronoun we. She mainly distinguished between inclusive-we and exclusive-we, as other researchers have done before and after her (Biber et al., 1999; Haas, 1969; Kuo, 1998; Levinson, 1992; Pennycook, 1994; Spiegelberg, 1973). Inclusive-we includes the hearer or audience in its reference scope, whereas exclusive-we excludes the hearer or audience from the reference. For Rounds, however, this classication was not sucient and she added three more types of we, according to their referent (Rounds, 1987a, pp. 1819):  we in contexts in which I is more specically marked. For example, We said that. . . The teacher is repeating some previously explained information.  The students are the actual sole referents of we. For example, I want to look at some of the problems we had for today. . . The students had the problems as homework.  we which has as its actual referent anyone who does calculus. Rounds considered that this we could be substituted by one (Rounds, 1985, pp. 54 55). For example, We [mathematicians] call that number. . . This would be considered an example of exclusive-we, since the hearers, the students, are not included in the reference scope. The third part of Rounds research consisted in dening discourse contexts for the dierent types of we (1987, pp. 1920). She provided Table 1,which shows the referents, the discourse contexts and other alternative pronouns.
Table 1 Referent and discourse contexts of personal pronouns (Rounds, 1987a) Referent Teacher Students Students and teacher Mathematicians and teacher Anyone who does calculus Pronoun I, we You, we I, you, we I, we I, you, we Discourse contexts Reporting previous remarks; announcing future actions Referring to student responsibility; admonishments Working with specic examples; announcing future actions Naming; dening Working with mathematical procedures

1 In Rounds research, success in teaching is rated by the department supervisor, teachers are either good or problematic based on rst-hand observations, end-of-semester evaluations, and the number of student complaints led with his oce (1987a, p. 15).

I. Fortanet / English for Specic Purposes 23 (2004) 4566

49

My rst intention when starting this research was to study the use of personal pronouns in spoken and academic English since they may represent a diculty for the understanding of international students who are not used to academic speech, and who may not understand at once the referent and function of certain uses of pronouns. Since Rounds work, no other studies on pronoun use in classroom speech have been carried out of which I am aware. Thus, her research was considered a starting point for the present study, which focuses on the use of the pronoun we. She used a corpus of ve lectures recorded for the research leading to her PhD dissertation, and her results showed a very frequent use of the pronoun we, which exceeded the other pronouns, I and you. Today a corpus of academic speech much larger and more multidisciplinary than that used by Rounds is available, the MICASE, from the same university where she compiled her own corpus, the University of Michigan. Though this circumstance gave me the opportunity of comparing her results with mine, my intention was also to conduct a deeper study taking advantage of the resources of corpus linguistics now available. To this end, the following research questions were established: 1. What is the frequency of we in academic speech today, compared with the use of other pronouns? Is there a signicant dierence if the corpus is restricted to one discipline such as mathematics? 2. In which linguistic contexts does we appear? 3. What referents can be found for we in a corpus of spoken academic English? 4. What are the discourse functions of we? Though I tried to answer all four questions, it was questions 3 and 4 that received most attention.

2. Method As said before, MICASE was used as the corpus of the present research. MICASE is, as is stated in its web page, a research database open to the world (Simpson, Briggs, Ovens & Swales, 1999), since it can easily be assessed on line on the Internet. It consists of a corpus of approximately 1.7 million words transcribed from a variety of speech events at the University of Michigan at Ann Arbor. Recordings and transcriptions began in February 1998 and are still going on. The speech events transcribed are classied into classroom and non-class events. Among the former there are lectures, discussion sessions, lab sessions, seminars and student presentations. Among the latter, there are advising sessions, colloquia, dissertation defenses, interviews, meetings, oce hours, service encounters, study groups, tours and tutorials. The speech events are related to a variety of academic departments in the biological and health sciences, the physical sciences and engineering, the social sciences and education, and humanities and art. There is also a wide range of participants that represents almost the entire university.

50

I. Fortanet / English for Specic Purposes 23 (2004) 4566

As established by the research questions, the rst focus of this research was the frequency in the use of we. For this quantitative study, I used Wordsmith Tools (Scott, 1996) to analyse one of three sub-corpora I established for this studyCorpus A, which included all lectures and colloquia in the MICASE corpus, up to 75 speech events with a total of 770,353 words. The frequency of we, and related words (us, our, us, lets) was compared with the frequency of you and I and their related words. Since the results obtained using this corpus were dierent from those found by Rounds, a second search was carried out with corpus B. Corpus B consisted of a lecture, a colloquium and a study group transcript, with a total of 28,477 words. Although it would have been preferable to use a corpus of only lectures, these were the only three speech events in the MICASE related to the discipline of mathematics, which was the one selected by Rounds for her corpus. The linguistic contexts of we were also considered as relevant for this research. Using again Wordsmith Tools, the repetition of collocation clusters was analyzed in Corpus A. The results were also employed later to support the discourse functions of some uses of we. The two most important parts of the research were the search for the referents and the discourse functions of we. I immediately realized that the application of Wordsmith Tools to Corpus A would not help me with these two variables. A thorough qualitative analysis was needed, which required a careful reading of a more limited corpus. For this purpose, Corpus C, with a total of 40,986 words, was created. It consisted of four lectures, three of them addressed to students with 40 or more people in the audience, and the fourth addressed to a more limited number of faculty members. Here is a brief description of the four lectures: Lecture 1: Education Colloquium (duration: 52 min) (9119 words). In this lecture the audience was formed by a number of faculty members and, though the rst part of the transcript is a monologic lecture, the second part is a colloquium with the participation of eight of the faculty members in the audience. Lecture 2: Intro Anthropology Lecture (duration:74 min) (11,512 words). This lecture was delivered in an auditorium to 400 students. It was not followed by a colloquium, but there were some questions asked by one student in the middle of the lecture. Lecture 3: Japanese Literature Lecture (duration: 44 min) (8534 words). This lecture was attended by 60 students, and up to 6 students from the audience asked questions and made remarks during the lecture. Lecture 4: Medical Anthropology Lecture (duration: 69 min) (11,821 words) This lecture was attended by 40 students. In this case the speaker did not keep the oor, but let some students in the audience participate, not only with questions at the end but with remarks and comments during the lecture. Most lectures in US universities are followed or interrupted by the participation of members of the audience. Audience intervention changes a monologic mode of language into a dialogue. It is important to know this, since it may aect directly the use of pronouns, their reference or their function. In the description of the lectures I

I. Fortanet / English for Specic Purposes 23 (2004) 4566

51

pointed out the participation of other speakers in all the speech events in this corpus. The primary mode of speech in all lectures is monologic, although dialogue is also present in all of them.

3. Results 3.1. Frequency A search using the Wordsmith Tools Concordancer produced the results in Table 2. Rounds results were added to the table for comparison. However, her results were only given in number of occurrences and it was dicult to establish a comparison between two corpora so dierent in length. Since the concordancer provided us with the frequency rates in the MICASE ( per thousand words), this rate was also estimated for Rounds corpus.2 The third person singular and plural were excluded from this analysis since the referent can very rarely be the speaker or the hearer. Even though the concordancer could not make a distinction between the uses of you as subject and object, it seems clear from the data that the use of the pronouns I and you individually and collectively nearly double that of we in the MICASE. The occurrences of the possessive pronouns are more or less
Table 2 Occurrences and frequency of personal pronouns Pronouns MICASE Occurrences First person I me my total we us lets our total Second person you (sub & obj) your total Frequency (per 1000 words) 17.9 1.53 1.6 21.03 9.7 0.8 0.8 1 12.3 Rounds corpus Occurrences Frequency (per 1000 words) 11.5 1 0.1 12.6 34 1 3.5 0.6 39

13,827 1180 1244 16,251 7450 610 644 785 9489

301 26 2 329 907 35 92 18 1052

16,000 1664 17,664

20.7 2.1 22.9

335 3 338

12.5 0.1 12.6

Since I knew the number of words in Rounds corpus, I estimated the frequency rate, that is, number of occurrences per thousand words. In this way, all data were comparable.

52

I. Fortanet / English for Specic Purposes 23 (2004) 4566

correlative to the use of the personal pronouns as subjects. However, in the comparison of the object pronouns, us in its full and abbreviated form is relatively more frequent than me. Although we does not seem to be so frequent in academic speech as the other pronouns, there were still sucient numbers to make it worth analysing. The results from the MICASE in Table 2 dier most noticeably from those of Rounds, in that the occurrences of we more than triple those of I and you. In view of these results, a second search was conducted using Corpus B, a selection of speech events related to the mathematics, the discipline of Rounds corpus. Table 3 shows the results. In this reduced corpus the occurrences of I exceed considerably those of we and you. These results are dierent from the ones found in Corpus A, probably due to the relatively small size of the corpus. Comparing them with Rounds results, it can be observed that apparently the use of we is higher than that of you, as she also found. However, it is the frequency of I that is most noteworthy in this table, which more than doubles the rate of both we and you, and is also three times the number of Is found by Rounds. We cannot, then, state that the results in the search of Corpus B support Rounds research. It is perhaps not the disciplinary and instructional character of mathematics that encourages a greater employment of we, and an explanation for Rounds anomalous resultsin terms of the bigger MICASE pictureneeds to be found elsewhere. As some part of the research was carried out using Corpus C, the occurrences and frequency rates of we were also observed for the four lectures in this corpus. Table 4 shows the results.

Table 3 Occurrences and frequency of personal pronouns in mathematics lectures in the MICASE Pronouns First person I me my total we us lets our total Second person you (sub & obj) your total Occurrences Frequency (per 1000 words) 30.4 2 1.5 33.9 14.8 0.8 3.1 0.6 19.3

812 61 46 919 452 26 95 20 593

339 3 342

11.1 0.09 11.2

I. Fortanet / English for Specic Purposes 23 (2004) 4566

53

Comparing these results with those obtained in Corpus A from the MICASE, which showed an average of 9.7 per thousand words, the minor deviations in lectures 3 and 4 are insignicant. 3.2. Clusters search The second aspect of our quantitative research on the use of pronouns in academic lectures was to observe the presence of clusters associated with we. Table 5 shows the rst 30 clusters Wordsmith Tools provided. We appears very frequently with have preceded and followed by other words: we have to, we have a, that we have, we have the, so we have, we dont have, and we have, to total 512 occurrences. It is also common to nd the pronoun we with talk: we talked about, well talk about, were talking about, we talk about, were gonna talk, making a total of 274. Other clusters that appear near the pronoun we include the verb look, know, can/ be able to, need, going to, etc. Next, these clusters will be related with the discourse functions of we in the four lectures of Corpus C.
Table 4 Occurrences and frequency of we in Corpus C Occurrences Lecture 1 Lecture 2 Lecture 3 Lecture 4 86 111 92 99 Frequency rates (per 1000 words) 9.4 9.6 10.7 8.4

Table 5 Clusters associated with the pronoun we N 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Cluster we have to we have a we talked about we need to were going to how do we we have the that we have so we have well talk about were talking about and then we that we can we dont have we look at Frequency 106 98 85 77 76 74 71 70 65 57 55 48 47 44 44 N 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 Cluster we know that what do we you know we and we have i think we and we can we talk about were gonna talk a little bit a lot of and so we and then well were trying to gonna talk about so if we Frequency 43 42 42 41 38 37 37 35 33 33 33 33 33 32 32

54

I. Fortanet / English for Specic Purposes 23 (2004) 4566

3.3. Referents and discourse function Beyond the analysis of clusters, a thorough study of the four lectures in corpus C was undertaken in order to establish, in the rst place, the reference scope we can have and the discourse functions in which it appears. However, before presenting the results of this analysis, I would like to point out two main diculties found in this analysis. The rst is that in all four of the lectures examined there were a number of occurrences of we that could not be classied. Most times these occurrences were followed by pauses and a change of mind of the speaker, who then used another word interrupting the discourse. These unclassied occurrences ranged from 5 to 7 in the four lectures. The second problem when analyzing the transcriptions of the lectures in corpus C was to identify reported speech. It is found in many occasions in the form of free direct speech; that is, the speaker quotes the words as spoken by another anonymous person, but without advising the hearer s/he is going to do so. At some other times, the speaker adds a short reporting phrase just before or after the citation, such as this is what one person explained to me or I heard this quote. (1) this is what a faculty member told me. we hardly ever discussed anything3 (Lecture 1) When there is an absence of reporting phrases, the presence of quoted speech can be noticed by a shift in the use of personal pronouns towards I or we, in the words of the reported speaker. (2) theyve also, individually, changed courses. added new readings, brought in examples from dierent countries, they tried to make it more international. and then secondly their pursuing internationalization. part of it because if we didnt have an international program wed be ranked much lower. (Lecture 1) This type of reported direct speech provides vividness to the speech and can also be considered as an informality feature, but it requires an appropriate context and special intonation, information which is not provided by transcripts. Table 6 summarizes the results obtained for referent and discourse function in Corpus C, which will be discussed next. Though I have just used one table to summarize all the results, these will be discussed separately. 3.3.1. Referents As can be seen in Table 6, the highest number of occurrences of we was found in reported direct speech with the same referent, larger group of people (including the reported speaker). The frequent presence of this pronoun in this peculiar type of reported speech can be attributed to the lack of identity of the speaker who cannot speak for him/herself and must then speak as a representative of a group. In the
3

Underlining means reported direct speech.

Table 6 Referent and discourse function of we in four lectures from the MICASE Reference Larger group of people (including speaker+audience) Occurrences 78 No. of lecture 1,2,4 Inclusive we/ Exclusive we i Discourse functions  Representation of other communities of people the speaker belongs to (Western civilization (as opposed to they=African tribes), humans, etc.)  Guide throughout the speech event  Presentation of the situation  Guide throughout the course/ discipline  Joint deduction  Presentative have or get (=there is)  Guide through the speech event  Guide throughout the course/discipline  Recount of the research process  Clarication  Guide through the speech event  Representation of speaker and Masai people. The speaker tries to express his involvement in the situation described.  Indication of belonging to a scientic community.  Generalizing without losing the audiences involvement

I. Fortanet / English for Specic Purposes 23 (2004) 4566

Speaker+Audience

we for I

we for you (audience) Speaker+other people

Total: 78 27 4 12 4 24 Total: 71 41 4 22 3 Total: 70 13 Total: 13 3

2,3 1 2,3 2,3 3,4 1,2,3,4 4 1 2 2,3,4 2

i i i i i i i e i i e

2 Total: 5 4 Total: 4

we for indenite you or one

(continued on next page)

55

56 I. Fortanet / English for Specic Purposes 23 (2004) 4566

Table 6 (continued) Reference we for they Occurrences 1 No. of lecture 2 Inclusive we/ Exclusive we e Discourse functions  Representation of lm makers. The speaker tries to express his involvement in the situation described.  Identication with toy manufacturers.

2 Total: 3 Reported direct speech Larger group of people (including the reported speaker)

96

1,2,3,4

 Reported speakers are not often singled out as a person, they are most often a representation of a group of people.  sympathy parent-child

we for you

Total: 96 1 Total: 1

I. Fortanet / English for Specic Purposes 23 (2004) 4566

57

following example, the speaker wants to illustrate the doctors reaction to the Inuit native womens behavior when they were about to deliver a baby, and did not want to move from their isolated villages. (3) . . .and when folks came in for their check up and they were about three (few) weeks away from their due date (people) would say now, you know we dont want you just waiting too long and going into labor so (as) we have to deliver you here (Lecture 4) In the example, the introducing speaker says people in a low voice (signaled by parentheses) to indicate the referent of we though from the context it can easily be reduced to doctors. This use of we is always exclusive, since there is no logical link between the reported speaker and the audience. Citing or quoting anonymous speakers seems to be often used as a linguistic device of hedging, since, in this way, it is other people, and not the speaker, who express their opinion. Moreover, cited or reported words do not necessarily have to be the real words said by the reported speaker or it may even be possible that the reported speech event never happened; what the speaker presupposes the hearer knows and accepts. The referent of we as larger group of people (including the reported speaker) found in reported direct speech can also be related to the second most frequent use of we in the MICASE, larger group of people (including speaker+audience), with 78 occurrences. In contrast to the previously reported use, here we includes the two participants in the speech event together with other people. In the following example, the shift made by the speaker from they to we makes the referent very clear to the hearer. (4) . . .humans have come up with m- terms of how to, acquire those things that they need, that we need. . . (Lecture 2) The speakers intention in this case is to involve the audience in what s/he is saying. There is another referent of we much less frequent than the previous ones (only 5 occurrences), which includes a group of people and the speaker, but excludes the audience. Unlike the results obtained for the previously mentioned use of we, the intention of the speaker seems to be here a detachment from the audience, a distancing between the experts, including the instructor-speaker, and those who are still learning, the students-audience. (5) . . .you always see comparisons of infant and uh and and perinatal mortality used as indicators of uh just how advanced a country or a community is we use uh these sort of statistics, as ways to talk about how good things are uh in dierent places,. . . (Lecture 2) A second group of referents of the pronoun we includes all those in which we is employed referring either to the speaker or to the audience or to both of them, but

58

I. Fortanet / English for Specic Purposes 23 (2004) 4566

excluding other people. I is a very frequent referent of we in the MICASE corpus C. The speaker involves the students in actions s/he can only do, as can be seen in the following example. (6) . . .it is because were going to be talking about it today. . . (Lecture 4) It is only the teacher who presents and explains the new topic, not the students. Most of the occurrences of we with I as a referent are inclusive; that is, the speaker tries to involve the audience in what s/he is doing. However, in Lecture 1 there is a high concentration of wes with this referent that are exclusive of the audience. The speaker is talking about some research she carried out some time ago, and she primarily uses I during the lecture, but once the colloquium starts she starts using both I and we, becoming almost exclusively we at the end of the colloquium. The shift can be observed in the following example: (7) S1you know i did not ask faculty about that so i couldnt ask that uh i couldnt answer that in this survey < PAUSE DUR=:0400 > i wish i_i dont remember all the questions i know_, im not sure we asked that specically In this example, she starts by using I and nishes by changing it into we, even though the verb remains the same. At the end of the colloquium (two last interventions) I has almost disappeared (three occurrences), and we has replaced it (nine occurrences). I contacted the speaker by e-mail4 and asked her what she thought could be the reason for this change of personal pronoun in her speech. She replied that it was due to the fact that she had been asked about some part of the research she had carried out with a team of people. However, I believe that although this may be the reason for some of the wes in the colloquium, it does not explain examples like the one provided above. I think this shift from I to we may be due, on at least some occasions, to an unconscious resource of hedging, that is, the speaker is trying to protect herself from the questions and opinions of the other faculty members in the audience by using a royal we. Another very common referent of we includes both the speaker and the audience. Obviously, all the occurrences of we with this referent are inclusive. The intention of the speaker is to express the actions shared by him/herself and the audience in the speech event, as in example (8): (8) . . .when we read Mori Ogai, next were gonna read about that a little bit. . . (Lecture 3) You can also be identied as the sole referent of we when the action expressed can only be carried out by the students.
4

Personal communication by e-mail (2 February 2002).

I. Fortanet / English for Specic Purposes 23 (2004) 4566

59

(9) what are we missing? Somebody, did anybody think theres something missing? (Lecture 3) Here the teacher asks the students what they are missing. The logical pronoun should be you, but the teacher has chosen we for his/her question. Sometimes it is dicult to distinguish in the transcribed text if the referent of we is the speaker, the speaker and the audience, or just the audience. It is the verbs used by the speaker that give the clue to the referent as in the following examples from Lecture 2: (10) . . .last, week we also heard about a Masai group in east Africa (Lecture 2) (11) . . .were gonna see that right now (Lecture 2) In example (10), it is the students who hear what the teacher says. In example (11), both speaker and audience can see what is coming next. There are two other uses of we that have you as a referent, which are more specic. In four occurrences we substitutes an indenite you or one. For example: (12) . . .they start doing this again, which would be like if we were to cough extendedly or hiccup extendedly (Lecture 2) The meaning is exactly the same if we is substituted by you or one. The second is found only once in reported direct speech, but it responds to a linguistic phenomenon commonly found when speaking to children in a paternalistic way, often to scold them. The adult avoids using you to make the sentence sound less strong; moreover, with the same aim, instead of an imperative a question is often used. (13) . . .we want to be safe now (dont we). Put on that cape and out the window you go (Lecture 4) The speaker is mocking the exaggerated risk warnings on toys and presents what parents would never tell their child to do, to put on a toy Superman cape and y out of the window. Finally, we can also refer to they when the speaker tries to involve the audience and him/herself with a situation not related to either of them. In example (14) the speaker is talking about the absurd warning labels on some toys. He illustrates the explanation with an anecdote about a plastic sledge from Toys R Us. When he has enumerated all the warning instructions in the label, he says: (14) . . . have we eliminated everything now? (Lecture 4) From the context, the audience can easily understand that we refers to the people in the company Toys R Us. However, it can also be understood as a generalization

60

I. Fortanet / English for Specic Purposes 23 (2004) 4566

which involves the whole of US society, referring to their behavior, and to their fear of risk. Looking again at Table 6, it indicates that 203 out of 328 (62%) of the occurrences of we are inclusive-that is, they include the speaker and the audience as their referent-whereas only 125 (38%) are exclusive, most of them in reported direct speech. 3.3.2. Discourse functions Though they may seem very dispersed in Table 6, the discourse functions found in corpus C from the MICASE can be grouped in two categories: representation of larger groups of people and metadiscourse. 3.3.2.1. Representation of groups. The most frequent category of discourse functions is that related to representation of groups, which includes the uses of we excluding the audience found in reported direct speech, and we with the referent speaker and other people, larger group of people, including the reported speaker, the use of we meaning speaker and audience, and we for they. The groups of people represented range from very broad ones, such as people in general: (15) . . .lots of people talk of risk risk is is a very colloquial part of the way we talk about, things that have to do, with health and illness (Lecture 4) to more restricted ones, such as a group of Masai: (16) . . .so we go onto the set, and they had kept the Samburu (Lecture 2) The speaker provides an explanation for this use of we: . . . I w-I was with them because I was the interpreter. 3.3.2.2. Metadiscourse. The second main category could be labelled metadiscourse functions, as all of them refer internally to the speech event in which they are found. The most frequent of these functions is guide through the speech event. The teacher informs the students about what is coming next, how what s/he has just said relates to what s/he is going to explain, etc: (17) . . .in terms of reputation we already talked about (Lecture 1) With this function we can also have as a referent the speaker and the audience: (18) . . .the idea of making a living will come up again and again today. And um, well think about what that means to make a living (Lecture 2) and you, including the audience: (19) . . . there are lots of articles weve already read in the course pack (Lecture 4).

I. Fortanet / English for Specic Purposes 23 (2004) 4566

61

Another metadiscourse function, which can also be found for we with I as its reference, as well as speaker and audience, is that of guide through the course or discipline. The speaker takes the role of teacher who guides students in their learning a discipline. S/he relates what s/he is saying with what has been explained before or will be explained in the future, and also with other subjects, but involving students in it, in this way avoiding an authoritarian representation of him/herself. This function is reected in the clusters search, in which a high frequency was found for the use of we know that (see Table 5): (20). . .we know, we know from history that. . .(Lecture 2) There are other discourse functions of we in this category that were only found with one referent. In Lecture 1, the speaker recounts a research process using we for I, but excluding the audience: (21) . . .we asked them to put their department down [. . . and we got a lot of negative responses (Lecture 1) Clarication is a minority function of we with I as a referent: (22) . . . when we say make a living were generally thinking about (Lecture 2) Another metadiscourse function is the presentation of a situation. In the occurrences with this function, we has the referent speaker and audience: (23) . . .its a timely topic to be looking at what were doing with the international students, who we do currently enroll (Lecture 1) It should be born in mind that the audience in this lecture is formed by academics, not students. The speaker presents the situation before going on with the recount of the research process. Joint deduction is found as a metadiscourse function with I as a referent in Lectures 2 and 3: (24) so now we think we know reality. okay we have, we thought we knew (Lecture 3) In all the examples found of this function we is preceded by so. So seems to have two important functions in academic speech, one is what Swales and Malczewski (2001, pp. 147, 157158) have called new-episode ags, that is, a linguistic resource used by academic speakers to recover the attention of their hearers. The second function of so is to express some kind of summative evaluation or the onset of summation, as also noticed Swales and Malczewski (2001, pp. 156158). This second function is more relevant for the present research, since so is in some cases accompanied by we followed by a verb. This is supported by the high number of clusters of so we have (65) that were found in Corpus A, as

62

I. Fortanet / English for Specic Purposes 23 (2004) 4566

shown in Table 5, though have is not the only verb that appears with we with this function: (25). . .so, we have large-scale agriculture here, where families. . .(Lecture 2) (26). . .so, because of [. . .], we nd that foragers cannot,. . .(Lecture 3) Finally, there is a function in this metadiscourse category that calls for special attention: we is used very often, especially in Lecture 3 with the verbs have and get with a presentative function, that is, the speaker uses this resource to introduce something new to the audience. In all cases, we get or we have can be substituted by there is/there are. The clusters search support these ndings though only in relation to have, since so we have comes ninth in the list of the most frequently used clusters: (27). . ., and we get [ there is] this whole list of names and theyre all equated with the Japanese syllabary,. . .(Lecture 3) (28). . .and that was a problem because we had [there were] three authors for this particular play. . .(Lecture 3) (29). . .you know we do have sort of [ there is ] this um, needless gore. . .(Lecture 3) Biber et al. (1999, pp. 955956) note the parallelism between some existential clauses with there is/there are and others with have but no reference is made to the verb get and the presentative meaning of we have/get. Many of the occurrences of we with this function also have the demonstrative this in the near context. However, even though the presence of this pattern is supported by the ndings of clusters, the high concentration in Lecture 3 suggests it may just be a feature of the personal style of the speaker, rather than a general trend, though this point would need further research.

4. Discussion This study explored the use of the pronoun we in lectures. The results showed that, though this is not the most frequent pronoun, it still has an important presence in spoken academic English. We has a wide range of referents that go from large groups of people to merely the speaker. The discourse functions of this pronoun are also numerous and they seem to t into two categories: representation and metadiscourse. In terms of frequency, the Wordsmith Tools Concordancer indicated that we is only used half as many times as other pronouns such as I or you in todays academic English. The corpus analyzed consisted of 770,353 words and was selected from the MICASE. These results are dierent from those obtained by Rounds

I. Fortanet / English for Specic Purposes 23 (2004) 4566

63

(1985, 1987a, 1987b) fteen years before. In order to look for an explanation for this substantial dierence in relation to the discipline, a reduced corpus of academic talk related to mathematics was also analyzed. A possible reason for the high percentage of wes found by Rounds, which cannot be validated with the present research, can be related to recent tendencies observed in academic language. It may not be entirely co-incidental that during the last 20 years, the use of I has become more acceptable in research articles in a number of elds, very often substituting for royal wes (Hyland, 2001; Chang & Swales, 1999), and the same kind of substitution may be happening in spoken academic speech. As far as referents are concerned, most occurrences of we refer to a large group of people, of whom the speaker is the representative or spokesperson. Rounds identied this referent when used by the teacher, excluding the students, as an indication of belonging to a scientic community, and when referring to metadiscursive functions and including the audience. However, the most frequent use of we as representative of large groups, such as humans, or members of the Western civilization, was not identied. Nothing is mentioned either about the use of we in anonymous direct reported speech. Other referents already found by Rounds are I (the speaker/ teacher), you (the audience/ students) and speaker+audience (teacher and students). This taxonomy has been completed by some other referents found in the present study: speaker and other people, indenite you or one, and they. As can be observed from this variety of referents found in academic speech, it cannot be assumed by the listener that we represents only the speaker and the audience. As Biber et al. (1999) pointed out, it often has a vague meaning that has to be gured out by the hearer from the context. Many researchers have already distinguished between exclusive we and inclusive we, including Rounds, but no data had been provided comparing the frequency rate of one and the other. The results of this research seem to suggest that inclusive we is more often used in academic speech than exclusive we. Though more evidence would be needed to make an assertive statement, this result may suggest that teachers use inclusive we in an eort of co-operation with the students and exclusive we to create a distance between the speaker-teacher and the audience-students. Since inclusive we prevails over exclusive we, it could be said that we is more often used as a co-operative than as a distancing device nowadays in academic speech. With regard to discourse functions, two categories have been identied in this research: those related to the representation of a group, and those related to metadiscourse. The discourse functions of representation are closely linked to the referent of representative or spokesperson of a group. As already noted about this referent, the discourse functions in this category, despite being the most numerous, had not been identied in previous research, such as that carried out by Rounds. Many of the occurrences of this discourse function happen in direct reported speech, which seems a common resource in academic speech. Considering the fact that in all these occurrences it is exclusive we that is found, it could be stated that it helps to support the distancing eect produced by the use of direct reported speech.

64

I. Fortanet / English for Specic Purposes 23 (2004) 4566

Regarding the second category of discourse functions, they all relate to the situational speech event, the lecture; or to the global speech event, the discipline:         guide throughout the speech event guide throughout the discipline presentation of the situation joint deduction clarication generalization recount of a research process presentative have or get

The rst two discourse functions, those of guide, are the most common in all lectures analyzed, followed by presentation of the situation, joint deduction, clarication, and generalization. Then, there are two functions that are also very frequent but are only found in one lecture, recount of a research process and presentative have or get which could account for a personal style of the speaker. Only further research would support a generalization of these ndings. Metadiscourse has also been identied in academic speech, even in previous studies using the MICASE as a corpus. Mauranen (2001, pp. 169172) includes this discourse function in what she considers discourse reexivity, or metadiscourse, which is characterized by organizing, describing and commenting on the ongoing discourse, and by guiding the hearer in the interpretation of the discourse, even though her analysis deals with reexivity, and not with pronouns. This suggests that metadiscourse may identify an important part of classroom speech, with specic elements of discourse to mark this function. In the introduction to this paper, the writer roles identied by Tang and John (1999) were described. Although this study was made with a corpus of academic writing, and dealt with the pronoun I, I tried to apply Tang and Johns results to the corpus of this research, and found that a correspondence may be established between some discourse functions and the roles they identied:  Societal role: identities inherent to a person. This can be related to the function of representing a group of people. Depending on the group of people represented, the speaker can adopt one or another societal role: American, human, faculty member, etc.  Discourse role: identities that a person acquires by participating in a particular discourse community (lawyer or client in a legal discourse community, teacher or student in an educational environment, etc.). The speaker in academic speech acquires the discourse role of teacher when s/he adopts the discourse function of guide throughout the discipline or course.  Genre role: specic to a particular genre within the discourse community. It is adopted by the speaker with the discourse function of guide throughout the speech event.

I. Fortanet / English for Specic Purposes 23 (2004) 4566

65

In conclusion, we is rather frequently used in academic speech, though not so often as other pronouns such as I or you. We can be found with a wide variety of referents and discourse functions and it is up to the speaker and hearer to negotiate which of them is used. This negotiation is usually helped by linguistic and extra linguistic clues. These clues, which are easily detected by a native speaker of English, can represent an signicant hindrance for a non native speaker of this language. Studies like the present one help to create a mapping of the use of English in the classroom not only in order to add to a description of the English language, but also to assist non native students in their academic objectives.

Acknowledgements I would like to thank Professor John Swales of the English Language Institute at the University of Michigan for his valuable feedback and helpful comments throughout the course of the research and writing of this article. I would also like to thank Dr. Jane Sunderland from Lancaster University, and two anonymous reviewers of this paper for their constructive criticism, which has improved and claried my views considerably. The research presented in this paper was funded by the Generalitat Valenciana (GV 00-065-9)

References
Biber, D., Johansson, S., Leech, G., Conrad, S., & Finnegan, E. (1999). Longman grammar of spoken and written English. London: Longman.. Brown, P., & Levinson, S. C. (1994). Politeness (1st ed.) 1978. Cambridge: Cambridge University Press. Chang, Y-Y., & Swales, J. (1999). Informal elements in English academic writing: threats or opportunities for advanced non-native speakers?. In C. N. Candlin, & H. Hyland (Eds.), Writing: texts, processes and practices (pp. 146167). London: Longman. Cherry, R. D. (1988). Ethos versus persona: self-representation in written discourse. Written Communication, 5(3), 251276. Flowerdew, J. (1994). Academic listening. Cambridge: Cambridge University Press. nez, R. (1998). The ancestors of an oral eld of discourse. In I. Fortanet, S. Posteguillo, J. C. Palmer, Gime : Publicacions & J. F. Coll (Eds.), Genre studies in English for academic purposes (pp. 297305). Castello de la Universitat Jaume I. Haas, M. R. (1969). Exclusive and inclusive: a look at early usage. International Journal of American Linguistics, 35, 16. Halliday, M. A. K. (1994). An introduction to functional grammar. New York: Arnold. Hyland, K. (2001). Humble servants of the discipline? Self-mention in research articles. English for Specic Purposes, 20, 207226. Kamio, A. (2001). English generic we, you, and they: an analysis in terms of territory of information. Journal of Pragmatics, 33, 11111124. Kuo, C-H. (1998). The use of personal pronouns: Role relationships in scientic journal articles. English for Specic Purposes, 18, 121138. Levinson, S. C. (1983). Pragmatics (1st ed.). Cambridge: Cambridge University Press. Mauranen, A. (2001). Reexive academic talk: observations from MICASE. In R. Simpson, & J. Swales (Eds.), Corpus linguistics in North America (pp. 165178). Ann Arbor: The University of Michigan Press.

66

I. Fortanet / English for Specic Purposes 23 (2004) 4566

Pennycook, A. (1994). The politics of pronouns. ELT Journal, 48(2), 173178. Rounds, P. (1985). Talking the mathematics thorough: disciplinary transaction and socio-educational interaction. Unpublished PHD dissertation, The University of Michigan. Rounds, P. (1987a). Multifunctional personal pronoun use in educational setting. English for Specic Purposes, 6(1), 1329. Rounds, P. (1987b). Characterizing successful classroom discourse for NNS teaching assistant training. TESOL QUARTERLY, 21(4), 643671. Scott, M. (1996). Wordsmith Tools. Oxford: Oxford University Press. Simpson, R. C., Briggs, S. L., Ovens, J., & Swales, J. M. (1999). The Michigan corpus of academic spoken English. Ann Arbor, MI: The Regents of the University of Michigan [http://www.lsa.umich.edu/eli/ micase/micase.htm]. Spiegelberg, H. (1973). On the right to say we: a linguistic and phenomenological analysis. In G. Psathas (Ed.), Phenomenological sociology (pp. 129156). New York: Wiley. Swales, J. M., & Malczewski, B. (2001). Discourse Management and New-Episode Flags in MICASE. In R. Simpson, & J. M. Swales (Eds.), Corpus linguistics in North America (pp. 146164). Ann Arbor: The University of Michigan Press. Tang, R., & John, S. (1999). The I in identity: exploring writer identity in student academic writing through the rst person pronoun. English for Specic Purpose, 18(1), 2339. Inmaculada Fortanet obtained her PhDr. in English Philology in 1993. At present she is a Lecturer at the , where she teaches English to English Philology, as well as Business Universitat Jaume I of Castello Administration students. She is one of the editors of Genre Studies in English for Academic Purposes. She has published several papers on Internet English and on academic written and spoken English in Spanish mo escribir un and international journals. She has coordinated the recent publication of the book Co n en ingle s, a handbook for Spanish researchers who want to publish in English. artculo de investigacio Her current interests are in written and spoken academic English.