pronoun use in novice l 1 and l 2 academic writing

This paper explores the use of pronouns in student academic writing in L1 Norwegian and L2 English. Three aspects of pronoun use are examined: general pronoun frequency, pronoun reference, and the uses of ‘I’ and ‘we’. Students of English are typically advised to avoid being explicitly present in their academic texts (Lysvåg & Stenbrenden 2014), while students writing in Norwegian are often encouraged to use first-person pronouns. However, the results show that both sets of students use explicit self-reference in similar ways. Frequent uses of ‘I’ and ‘we’ are as conductors of research and as guides or navigators assisting the reader through the text.

[2] insights int o self -represe ntation in academic writing Since students that are writing in Norwegian within the human sciences are often encouraged to use first-person pronouns in their writing (cf.Rienecker & Jørgensen 2013), while the opposite is the case for students of English (see e.g.Lysvåg & Stenbrenden 2014), we may expect to find a higher frequency of "I" and "my" in the Norwegian texts.This has already been shown to be the case in published academic texts in English and Norwegian (Fløttum et al 2006: 70-71).However, discourse communities may employ different conventions, and research writing involves both presenting oneself as an authoritative professional persona, demonstrating an individual contribution to a research community and constructing oneself as a plausible member of this community (Hyland 2001: 209).Hence, the work the writers do is more complex than merely adopting norms and standards of writing, and it may be difficult for students to deal with conflicting expectations from guidelines, supervisors and teachers.
Similar points are made by Ivanic (1998), who discusses issues of identity in academic writing and highlights the difference between researchers who treat academic writing as merely a fixed set of practices that students need to adopt, and researchers who recognize the social construction going on in academic writing.Ivanic argues that in studying writer identity, we need to adopt both a constructionist and an interactionist point of view, paying attention to the immediate interaction between real individuals as well as the norms of the cultures in which they are writing (ibid: 77).Ivanic chooses to place these norms within the notion of discourse communities, even though she notes that the concept may be problematic (ibid: 79).It has been shown that there are disciplinary differences in academic discourse (cf.e.g.Hyland 2001Hyland , 2012)).Research disciplines may be considered academic discourse communities, and the degree and type of self-mention in research articles is one of the features that varies between them.
We suggest that the student writers under scrutiny in the present study are members or aspiring members of a discourse community, but it is not immediately clear which discourse community that would be.In Norwegian universities the position of traditional humanities may be said to be on the wane, while teacher education has gradually been changed from a three-year program to a five-year master's program.Hence, the focus on teacher education in academic institutions has increased.Many academics belonging to traditional humanities fields such as English literature or linguistics are now teaching students in teacher education.While these academics may be said to talk and write within one discourse community, the students will meet several such in their studies: in the present case both a humanities (Norwegian or English) and a social science tradition, since a large part of the students' credits come from courses in pedagogy.
In our study, we also apply a contrastive linguistics perspective.Norwegian students writing in their first language may meet other traditions and other expectations than Norwegian students writing in their second language.Studies have shown that there are differences in self-representation across languages in academic writing.For instance, Sheldon (2009) explores 18 English and 18 Spanish research articles in Applied Linguistics and Language teaching.She reveals similarities in these articles, but to an even larger extent the differences in distribution of first-person forms are prominent.Drawing on the works of Ivanic (1998), Tang and John (1999) and Starfield and Ravelli (2006), Sheldon draws up six categories of the use of 'I' in the articles (referring to both singular 'I' and plural 'we'): "I as conveyor of general knowledge, I as a guide or a navigator, I as conductor of research, I as evaluator of previous claims, I as originator of claims and the reflexive I" (Sheldon, 2009: 253).She finds that 'I' as conductor of research and the reflexive 'I' are more prominent in the English research articles, while the Spanish articles have more examples of 'I' as guide or navigator and 'I' as evaluator of previous claims, and concludes that identity is constructed discursively in different ways in English and Spanish.
This leads us to a question that is of interest for the present study, namely whether writers transfer knowledge of academic writing from their first language to their second language.Kuhi, Tofigh, and Babaie (2013) investigated 40 research articles within computer engineering, written in English by 20 Iranian and 20 American researchers.The greatest difference was in the use of exclusive versus inclusive "we" in the articles.Exclusive "we" refers to the writers themselves, with discourse functions that signal their presence as researchers in one way or other.Inclusive "we" refers to either writers and readers or the discipline as a whole.According to Hyland (2001), involving the reader can be viewed as a manifestation of positive politeness, since it may acknowledge the reader as an equal.The American researchers used inclusive "we" to a larger extent than the Iranian researchers (Kuhi et al 2013: 43), but the authors provide no explanation for this difference.
The fact that self-representation is prominent in research articles within the field of computer engineering might be surprising, but it is not only the discipline that decides which practices are adopted.Pho (2008) investigated 30 article abstracts from three journals within applied linguistics and educational technology.Although the fields differ along several dimensions, the major trend identified with respect to personal pronouns was that there are differences between the various parts of the texts: for example, in all the abstracts the part of the texts that described the methodology distinguished itself by including no personal pronouns (2008: 246).Pho also suggests that theoretical papers will probably have other realizations of authorial stance than the empirical research articles he investigated.This also applies to the other works we have cited.Since our student papers are theoretical, this suggestion should be kept in mind.
[3] material and meth od The material for the present study comprises 16 student texts in total, eight of which were written in Norwegian and eight in English.The essays were written as part of a course on the philosophy of science, and focus on central aspects of positivism and social constructionism.The texts range in length from 2,000 to 2,500 words.
The texts were first searched for all instances of personal pronouns (see section 4.1 below), and then these were manually coded for reference in two stages.The first stage was what might be characterized as a data-driven approach, in that the analytical framework emerged from the data.The results of this first stage are discussed in section 4.2.The second analytical stage focused on the first-person pronouns 'I' and 'we' and was based on the framework outlined by Sheldon (2009), as described in section 2 above.During this stage of the analysis, we identified the following uses of 'I': 'I' as "conductor of research", 'I' as "guide or navigator", and 'I' as "originator of claim".Examples (1)-(3) illustrate these uses (relevant occurrences of 'I' are in italics). (1) With limited knowledge about both positivism and social constructionism, I decided to work my way through it.
(2) In this assignment, I will start with a presentation of positivism in general, followed by the most important directions within positivism; the early positivism and logical positivism.
(3) I have chosen two (out of five) of his statements, which I regard as central to this perspective.
In example (1), the writer reflects on the process of during which the paper was written.The fact that these papers are not based on empirical data means that the writers cannot be expected to carry out fieldwork or experiments as such, but in this example we can see the author reflecting on her own approach.

PRONOUN USE IN NOVICE L1 AND L2 ACADEMIC WRITING
[97] OSLa volume 9(3), 2017 Example (2) shows how the author may be present in order to guide the reader through the stages of the text.
In example (3) the writer claims that two out of five statements are the more central of the five, and in such cases the author is the originator of a claim.
The results of the second stage of the analysis are presented and discussed in section 4.3. 1 [4] results This section has been divided into three subsections.We will first look at the range and frequency of pronouns occurring in the material, and then we will look at the reference expressed by these pronouns.In the final subsection, we take a closer look at the reference expressed by the first-person pronouns 'I' and 'we', using the framework introduced in section 3.

[4.1] The range and frequency of pronouns used
Table 1 provides an overview of the range of pronouns found in the Norwegian texts.One thing is immediately clear from this table: the writers display a relatively large degree of individual preference both in terms of the pronouns they employ and in which ones they use the most.Some pronouns are used by very few of the writers, and with low frequencies.This applies to 'meg' ('me'), 'du' (singular 'you'), and 'en' ('one').In the mid-range of items we find 'jeg' ('I'), 'oss' ('us'), and 'vår' ('our').Then there are the two pronouns that are found in almost all the texts, and with a relatively high frequency in most of them: 'vi' ('we') and 'man' ('one').It should be mentioned here that Norwegian has two pronouns corresponding in meaning to the English word 'one', namely 'en' and 'man'.As is evident from Table 1, the writers tend to use one or the other.In texts 1 and 3 we find relatively high frequencies of 'en' and none or very few instances of 'man', whereas the opposite is true for the remaining six texts.
[1] Sheldon's categories for uses of 'we' are the same as those for 'I', so due to space limitations these have not been illustrated here.There are both similarities and differences in the range of pronouns used in English and Norwegian.Firstly, a similar pattern emerges with respect to the frequencies of the individual pronouns, in the sense that some are used infrequently by a few writers, while at the other end of the scale we find pronouns used by most of the writers, if not always very frequently.Briefly, the most infrequently used pronouns are 'my', 'your', 'us', 'own', and 'ourselves', whereas the most frequently used are the first-person pronouns 'I' and 'we'.There are also some differences in the types occurring in the two subcorpora, but some of these may be explained by the slightly different pronoun inventories in the two languages.For instance, Norwegian has an object form of the second-person [2] The texts range in length from approximately 2,000 words to 2,500 words (cf.section 3), so the measure of 2,000 words was chosen to avoid inflating the results.
singular pronoun ('deg'), and two pronouns corresponding to the English 'one' ('en' and 'man').What is perhaps noteworthy is that we find two possessive forms in the English texts, 'my' and 'your', and although Norwegian equivalents exist, the writers have chosen not to use them.On the other hand, in the Norwegian subcorpus we find the object form of the first-person singular pronoun ('meg'), and this has not been used in the English texts.It should be noted, however, that all of these items are used infrequently and by only a few of the writers.
We will now take a closer look at the frequencies for the three most frequently used pronouns, namely 'I', 'we', and 'one'.For convenience, the frequencies for these in both languages are presented together in Table 3.It is clear from Table 3 that 'I' is used in more English texts than in the Norwegian subcorpus, but that in those texts where the pronoun does occur, the frequencies are fairly similar in both languages, with the exception of text 5 in the English subcorpus, which has a much higher frequency of 'I' than any of the others in either language.It is perhaps slightly surprising that so many of the English texts contain 'I', given that students of English are advised to avoid being explicitly present in their texts (Lysvåg & Stenbrenden 2014).However, given the relatively low frequencies in most of the texts, we should probably not exaggerate the importance of this finding, although we note that this result is the direct opposite of the tendency identified by Fløttum et al (2006: 70-71; cf.section 2).'We' is used more frequently than 'I' in nearly all the English texts, and the [3] The column for the Norwegian 'one' contains the combined frequencies for the Norwegian equivalents 'en' and 'man'.
same applies to the Norwegian texts.Two aspects of the frequencies of 'we' are especially striking: the large degree of internal variation within each subcorpus, and the relatively speaking higher "peaks" in the Norwegian material, where texts 2 and 4 have a frequency of 23.7 and 52.8, respectively.The two highest frequencies in the English material are 15.3 in text 3 and 19.1 in text 5.At this stage, it is difficult to explain this difference between the two languages, but we will return to this when we explore the reference expressed by the pronouns in sections 4.2 and 4.3 below.Finally, Table 3 shows that 'one' is used much more frequently in Norwegian than in English, but that in the majority of those English texts where this pronoun occurs, it is used with a similar or higher frequency to those found in the highest-ranking (with respect to this pronoun) Norwegian texts.Interestingly, if we look at the "pronoun profiles" for the three English texts with the highest frequencies for 'one', we see that there are relatively few instances of 'I' and 'we'.This may be an indication that the students are aiming for a more impersonal writing style, but still rely on active sentences instead of, for instance, passive sentences where a personal subject could be avoided altogether.
We now turn to the reference expressed by the pronouns in both languages.
[4.2] Pronoun reference Table 4 presents the normalized frequencies for pronoun reference in both subcorpora.Note that the categories are those that emerged from the analysis of the data and were not pre-defined (cf.section 3). the pronoun frequencies given in Tables 1 and 2. For instance, it is hardly surprising that there are so few references to the reader of the texts, since we find very low frequencies for the second-person singular 'you'/'du' and related forms in both subcorpora.The same applies to the fact that the author's explicit presence in the text is more common in English than in Norwegian: we saw in Tables 1-3 that first-person singular pronouns are more common in the English texts than in the Norwegian.We may also note that references to the writer and reader (typically by means of 'we') are generally infrequent.We will return to these features in the next section, when we explore in more detail the uses of 'I' and 'we'.
When it comes to the category labeled "people in general", this is perhaps not something one would normally expect to find so frequently referenced in academic texts.However, as shown in the following examples, the students discuss how people experience the world, and the high frequencies for this kind of reference are therefore easily explainable. (4) The central question Berger and Luckmann attempts to answer is exactly how human activity leads to the production of what we percieve as "knowledge".
(5) Vi oppdager ny kunnskap gjennom opplevelsene vi har og de situasjonene vi står i. 'We discover new knowledge through the experiences we have and the situations we are in.' A similar explanation exists for the relatively frequent references to "researchers": given the essay prompts (cf.section 3) and the focus of the course, it is not surprising that the students have found it necessary to refer to the work of researchers. [

4.3] Uses of 'I' and 'we'
In this section, the focus is on the distribution in percent of the uses of 'I' and 'we'.The framework is based on Sheldon (2009; cf.section 3), but only the categories found in the present material will be discussed here.We will start with the uses of 'I', of which there are three in the present material: 'I' as "conductor of research", 'I' as "guide or navigator", and 'I' as "originator of claim".The boxplots in Figure 1  hough as the plot shows there is a lot of variation, and it should also be noted that four of the eight texts do not use 'I' in this way at all.The size of the material does of course limit the generalizability of these results, but it would seem that this is an area where further research on larger samples might support the preliminary conclusion that these results reflect cross-linguistic differences between English and Norwegian student writers, at least as far as this text type is concerned.We might tentatively suggest a similarity between our results for this feature and those reported by Sheldon (2009: 253; cf.section 2): Sheldon found that 'I' as the conductor of research was more common in English research articles than in Spanish articles.The second use of 'I' that was found in the material is 'I' used as a guide or navigator in the text, and Figure 2 shows how this type is used in the material.guiding the reader through the text.We may note, however, the slightly different profiles within the two groups: among the English texts we find two texts with no instances of this kind, and two texts where all instances of 'I' belong in this category.The remaining four texts are spread out across the scale.The Norwegian subcorpus, on the other hand, has four texts concentrated at the top end of the scale, and the remaining four have no occurrences of this use of 'I'.Again, it is clear that more material is needed in order to determine whether there is a contrastive difference in this usage of the first-person singular pronoun.Sheldon (ibid: 253) found more of this type of 'I' in the Spanish material than in the English material, but the great degree of corpus-internal variation in the present material does not permit us to compare our results for this feature with those reported by Sheldon.
The third and final use of 'I' that will be discussed here is 'I' as the originator of a claim.Again, there are indications that there may be a cross-linguistic difference at play: it seems that the students writing in English are more eager to present their own ideas about the subject matter than is the case for the students writing in Norwegian.Four of the students writing in English have a proportion of this type of 'I' that is above 40%, whereas two of the students writing in Norwegian use this kind of 'I' in between 15% and 40% of the cases, and one student has a clear preference for using 'I' in this way, with over 80% of the occurrences belonging to this category.It might be surprising that Sheldon's category "I as evaluator of previous claims" (ibid: 255) is not present in our material.After all, the students are primarily addressing previous claims in these theoretical papers.However, we may hypothesize that this is due to reluctance towards disagreeing with or criticizing well-known philosophers.
Although we cannot overlook the fact that the size of the samples may skew the results or cause important tendencies to be overlooked, it does seem fair to conclude that there are several indications of potential cross-linguistic differences in the way 'I' is used in English and Norwegian academic texts.It would therefore be fruitful to expand the material and thus broaden the basis on which conclusions may be drawn.
We now turn to the various uses of 'we'.Four such uses have been identified in the material, of which three are in accordance with Sheldon's categories (2009), and one had to be added.The three categories taken from Sheldon's study will be dealt with first.
Figure 4 contains the distribution of 'we' as the conductor of research.Given the discussion of the use of 'I' as the conductor of research in relation to Figure 1 above, it is not perhaps surprising that 'we' as the conductor of research is similarly more frequent in English than in Norwegian.What is more interesting is that there are indications that Norwegian writers use 'we' in this sense more often than they use 'I', and the same seems to be the case for English writers.It may be that using this more "inclusive" form is considered slightly more impersonal and therefore more in line with academic conventions than the unequivocally self-referent 'I'.
When it comes to the use of 'we' as a guide or navigator through the text, as illustrated in Figure 5, there are definite differences between the use of the first-person singular pronoun in this way and the use of the plural equivalent.If we compare Figure 5 with Figure 2 above, we can see that both the English and the Norwegian writers are less fond of using the plural pronoun in this way.In fact, it is clear from Figure 5 that the two subcorpora display a remarkably similar corpus-internal dispersion for this use of 'we'. Figure 6 shows the dispersion and proportions of 'we' used as the originator of a claim.There seems to be a fairly substantial cross-linguistic difference here, since none of the Norwegian texts contain this use of 'we'.However, note that only three of the English texts contain this usage, and so, despite the fact that the percentages are quite high in two of the three, we may well be dealing with individual preferences here.Once again, a larger sample is necessary in order to determine whether this truly is a contrastive difference between student academic writing in English and Norwegian.The final use of 'we' that will be discussed here is a type that was not included in the 2009 study by Sheldon referred to above.In the previous subsection, the use of pronouns to refer to "people in general" was discussed, and its prevalence in the present material was described as a result of the students wanting to include examples of how people experience the world.Since the students are also people inhabiting the world, it should not come as a surprise that they often use 'we' to express this meaning, and this resulted in the addition of an extra category that has been labeled "people in general in the real world".The frequencies and dispersion for this use of 'we' are presented in Figure 7.It is evident from Figure 7 that the Norwegian texts contain higher proportions of this type of 'we' than the English texts do when seen as a group: only two of the English texts have frequencies that are within the main Norwegian range of between 50% and 100%.There is one outlier among the Norwegian texts that does not contain any such instances of 'we', but in general the students writing in Norwegian can be said to favor this use of 'we' more than the students writing in English.We hesitate to label this a feature of student academic writing in general, however, since it seems to be so intimately connected with the topic of the texts that make up the material for the present study.
[5] concluding remarks Phos' suggestion that theoretical papers will probably have other realizations of authorial stance than the empirical research articles (2008: 247) seems to fit our small study.'We' referring to people in general in the real world is prominent in the present material, and research into empirical studies has not utilized such a category.The research cited in section 2 highlights language, disciplines or discourse communities and genre (theoretical or empirical paper) as aspects that might influence self-reference in academic writing, but it is reasonable to expect that the text topic will also influence this feature.
Our study shows a great deal of individual variation, among both L1 and L2 writers.Due to the small size of the sample, this may be entirely coincidental.On the other hand, it may also be the result of independent academic confi- dence, or a result of novice writers' attempts to conform to standards about which they receive conflicting guidelines.There may also be different reasons for the variation in the L1 and the L2 texts.Although both student groups have to learn that research writing includes presenting oneself as an authoritative professional persona, demonstrating an individual contribution to a research community and constructing oneself as a plausible member of this community (Hyland 2001), the L2 writers have the added difficulty of demonstrating foreign-language proficiency.Building on the theories of Kramsch (2009) about the subjective aspects of L2 learning, we may see this as a feature of being a social actor or developing a "self" in the L2.
To sum up, it seems that the results of the present study indicate several areas that might be fruitful areas for further research.Larger samples of student academic writing in both languages are needed to determine whether the tendencies hinted at in the present material are actually indicative of crosslinguistic differences, differences in discourse communities or whether they simply result from personal preferences of the relatively few writers included here.It would also be useful to include expert academic writing in English and Norwegian, as this would allow for a broader basis on which to identify contrastive differences.The inclusion of expert material in both languages would also make it possible to discuss the possible effect of transfer on the L2 student writers, and of developmental factors on both groups of students.
illustrate the dispersion of 'I' as the conductor of research in both subcorpora.The grey dots indicate the individual texts.As we can see, this use of 'I' is extremely rare in the Norwegian texts, with only a single outlier.It is more common in the English subcorpus, alt-

FIGURE 1 :
FIGURE 1: 'I' as the conductor of research (% of total instances of 'I')

FIGURE 2 :
FIGURE 2: 'I' as guide or navigator (% of total instances of 'I') This use of 'I' is clearly more frequent in both subcorpora than 'I' as the conductor of research, indicating that both groups of students are concerned with

FIGURE 3 :
FIGURE 3: 'I' as the originator of a claim (% of total instances of 'I')

FIGURE 4 :
FIGURE 4: 'We' as the conductor of research (% of total instances of 'we')

FIGURE 5 :
FIGURE 5: 'We' as guide or navigator (% of total instances of 'we')

FIGURE 6 :
FIGURE 6: 'We' as the originator of a claim (% of total instances of 'we')

FIGURE 7 :
FIGURE 7: 'We' referring to people in general in the real world (% of total instances of 'we')

TABLE 1 :
Frequencies per 2,000 words of pronouns used in the Norwegian texts 2If we then turn to Table2, we find comparable frequencies for the pronouns used in the English texts.

TABLE 2 :
Frequencies per 2,000 words of pronouns used in the English texts ref erences Fossum, Heidi.2011.The Writer and the Reader in Norwegian Advanced Learners' Written English: A corpus-based study of writer/reader visibility features in texts by Norwegian learners of English and native speakers of English.Unpublished master's thesis, University of Oslo.