The Acquisition of English Contrastive Discourse Markers by Advanced Russian ESL Students

The Acquisition of English Contrastive Discourse Markers by Advanced Russian ESL Students

By Jilani S. Warsi

Introduction

Since the late sixties, a considerable amount of research has been conducted in the field of second language acquisition (SLA). The complicated process of language learning has attracted continuing interest from researchers in English, linguistics, psychology, and education. Such interest has led to the emergence of second language (L2) studies as an area of professional emphasis within academic communities taking into consideration both teaching and learning perspectives. The field of SLA has become a vibrant field with a literature of its own, frequently using explorations in first language (L1) as a starting point.

While much work has been done in studying the acquisition of English morphology, syntax, and phonology by non-native speakers of English, there is a paucity of research on the acquisition of English discourse markers (DMs) by English as a Second Language (ESL) learners. This area is relatively unexplored as of yet in SLA research. In this study, I will focus on the acquisition of English contrastive discourse markers (CDMs), but, however, nevertheless, despite th is/that, in contrast, on the other hand, on the contrary, and instead , by advanced Russian ESL students. The purpose of this study is to examine the difference between native English speakers’ use of the above- mentioned CDMs and advanced ESL learners, and to determine how distant or close to the standard form the advanced Russian ESL learners are.

It is assumed that all languages make use of DMs or some such devices, which allow the display of utterance relations, although the repertoire of devices and their various functions varies from one language to the next. Since DMs contribute to coherence in discourse and therefore facilitate communication, it seems reasonable to suppose that inappropriate use of DMs in an L2 could, to a certain degree, hinder successful communication, or lead to a misunderstanding from time to time. Since many L2 learners do engage in interactive discourse, or ultimately aim to do so, they are responsible for signaling the relations of particular utterances to those which precede and follow, and therefore in terms of communicative competence, L2 learners must acquire the DMs of their target language (TL). It is plausible to suppose that those non-native speakers who are competent in the use of DMs of the L2 will be more successful in interaction than those who are not. Furthermore, it may be the case, following Ellis’s review (1996), that successful interaction can facilitate learning of grammar, and so there may be a reciprocal relationship between the acquisition of DMs and acquisition of grammar. It is for these reasons that the study of the acquisition of DMs in an L2 merits attention.

Discourse Markers

Discourse Markers (DMs), as described in Fraser (1997), “are lexical expressions such as those shown in bold in the following examples.”

(1) a) We were late in leaving home. Nevertheless , we arrived on time.

      b) It should fly. After all , we followed directions.

      c) It’s been a lousy day. The rain spoiled our picnic. Moreover, John didn’t come.

      d) A: I like him . B: So , you think you’ll ask him out?

      e) We ought to speak to Harry about that point. Incidentally, where is he today? (Fraser, 1997)

The past few years have seen continuing interest in the study of DMs. As quoted in Fraser (1997), several different scholars have labeled DMs as “cue phrases (Hovy, 1994; Knott & Dale, 1994), discourse connectives (Blakemore, 1987, 1992), discourse operators (Redeker, 1991), discourse particles (Schoroup, 1985), discourse signaling devices (Polanyi & Scha, 1983), indicating devices (Katriel & Dascal, 1977), phatic connectives (Bazanella, 1990), pragmatic connectives (Van Djik, 1985; Stubbs, 1983), pragmatic devices (Vande Kopple, 1985), pragmatic expressions (Erman, 1987), pragmatic formatives (Fraser, 1987), pragmatic markers (Fraser, 1988, 1990; Holker, 1991; Schiffrin, 1987), pragmatic particles (Ostman, 1989), semantic conjuncts (Quirk et al., 1985), and sentence connectives (Halliday & Hasan, 1976)” (Fraser, 1997). Fraser (1997) goes on to describe a DM as a

“lexical expression which signals the relationship between the discourse segment of which it is a part, S2, and the foregoing segment, S1. Each DM has a core meaning, but the meaning is not conceptual, such as is the case for the noun boy which denotes a young, male human, but rather procedural, where the DM signals how S2 is to be interpreted, given S1. For example, in (1a), where the S2 = “We arrived on time” and the S1 = “We were late in leaving home,” the DM nevertheless signals that we should interpret S2 as being in contrast with an expected implication of S1, in this case that we would be late in arriving.” (2)

It is obvious from the above definition that the main role of a CDM, according to Fraser (1997), is to establish a contrastive relationship between the two sequences, S1 and S2, which it connects. There are several other DMs that yield a contrastive interpretation of S1. Fraser (1997) calls this particular group of DMs Contrastive Discourse Markers, which will be discussed later.

The most detailed account of DMs comes from Schiffrin (1987), whose work illustrates the utility of investigating certain linguistic forms, which she calls discourse markers with regard to their role in bracketing units of talk, and thereby guiding the interpretations of utterances. Schiffrin (1987) states that these units of talk could be sentences, propositions, tone groups, and actions. Since markers can occur in initial and term inal positions, she defines brackets as “devices which are both cataphoric and anaphoric whether they are in initial or terminal position.” (p.31). The English markers which Schiffrin discusses are well, now, so, but, oh, because, or, I mean, and, y’know, and then. According to Schiffrin, discourse markers are “contextual coordinates for utterances: they index an utterance to the local contexts in which utterances are produced and in which they are to be interpreted.” ( p.36). “Local contexts” here refers to an utterance’s place within components or planes of discourse which co-occur and emerge in spoken interaction. That is, utterances occur simultaneously on different discourse planes.

Describing the different planes of discourse, Schiffrin distinguishes between two kinds of non-linguistic pragmatic structures: an exchange structure and an action structure. She notes that these are different planes of talk on which markers function.

1. Exchange Structure: According to Schiffrin, an exchange structure includes adjacency pair parts such as questions and answers, greetings, etc. The units of talk in an exchange structure involve two speakers switching their sequential roles in order to fulfill the mechanical requirements of talk imposed by one of the speakers; e.g. the hearer replying to a question asked by the speaker.

2. Action Structure: Speech acts under this category of non-linguistic structure are situated in terms of what actions are followed by persons in a particular situation. Shiffrin claims that these actions occur in a specific pattern and are predicable. In other words, they are not randomly ordered. They occur in “constrained linear sequences.” (p.25). These orders of occurrence show how people adhere to appropriate standards of interpersonal requirements of talk.

3. Ideational Structure: In contrast to the first two kinds of non-linguistic structures (exchange and action), Schiffrin views the units of talk within this structure as semantic, and considers an ideational structure linguistic in nature because they are “propositions with semantic content.” (p.26). She calls the units of talk within this structure “ideas”, and further explains the different relations between them: cohesive relations, topic relations, and functional relations. Establishing cohesive relations requires that the semantic interpretation of a clause follow the preceding clause. As evident by the term ‘topic relations’, these are dependent on the topics which speakers and hearers discuss. As compared to cohesive relations and topic relations, functional relations are concerned with the functional roles ideas play within a text.

4. Participation Framework: Schiffrin views the participation framework as pragmatic because of the relations of speakers and their intention, interpretation, and action. She stresses the fact that speakers and listeners can not only be related to each other by this responsibility to reciprocate in a talk, but that their relations are also influenced by what they are uttering.

5. Information State: This plane of discourse, according to Schiffrin, involves speakers and hearers using their cognitive capacities to organize and manage knowledge and meta-knowledge in the speech stream. She makes a distinction between knowledge and meta-knowledge here -- by knowledge she means what speakers and hearers know, and by meta-knowledge she means what speakers and hearers know about their knowledge and assume about each other’s knowledge.

Emphasizing that utterances occur simultaneously on different discourse planes, Schiffrin argues that the marker oh primarily marks information state transitions where a hearer displays recognition of familiar information, or else the receipt of new information. In addition, oh also plays a role in participation frameworks since it displays its user as a hearer who is “...an active recipient of information who acknowledges and integrates information as it is provided” (p.99). Finally, oh may play a role on the action structure plane of discourse by marking an utterance as an action, e.g., a clarification, which helps manage information state transitions.

Schiffrin asserts that a marker itself does not convey meaning, but rather contributes to the interpretation of utterance relations. In Schiffrin’s terms, markers “select” and “display” meaning relations between sequences of utterances. To demonstrate what these terms mean, consider the following example provided by Schiffrin:

a) Sue dislikes all linguists.

b) I like her.

The meaning relation between these two utterances is, on the surface, indeterminate. That is, one might select a contrastive relation between (b) and (a) (which could be marked by but before (b)), or alternatively are sultative relationship (which could be marked by so before (b)). Schiffrin (1987) argues that the interpretation of the meaning relation between these utterances in context would already be constrained by the background conditions. That is, the correct interpretation would be available based on the hearer’s knowledge of the speaker’s beliefs (e.g., that the speaker has a high opinion of linguists, or alternatively a very low one). Therefore, the addition of markers like but or so before (b) in this case would not determine the meaning relation between (a) and (b) but rather it would display a relation which was already inferable from the context.

In addition to displaying meaning relations, Schiffrin suggests that markers may also serve to display structural relations between utterances. That is, markers may display the identity of structural units in talk. Consider the next example (also given by Schiffrin):

a) I believe in fate.

b) I won the grand prize in a sweepstakes.

There are two possible interpretations of the meaning relations between (a) and (b) - (a) may be the cause of (b) or else (b) may be support for (a). There are also accompanying structural relations which are involved in that either utterance (a) could be the structural unit of a position with (b) as a unit of support, or alternatively (a) could be a unit of cause with (b) as a unit of result. Either of these interpretations is possible, but when markers like because or so are added, either interpretation of the meaning and structural relations between the utterances is selected and displayed to the exclusion of the other. Either of these interpretations was available without markers, and although one interpretation is likely to be preferred in actual context, the addition of markers makes more clear the relations between utterances.

However, as Fraser (1997) rightly points out, the notion that the hearer’s interpretation is derived from actual context does not always follow. To quote his example again:

a) We were late in leaving home.

b) We arrived on time.

The only possible interpretation the interlocutor can expect from (a) is a resultative utterance, such as “So, we arrived late.” However, it is the use of CDMs like but, however, and nevertheless, that makes the contrastive relationship between (a) and (b) clear. Discussing the discourse marker but, Schiffrin also notes that the use of but as a discourse coordinator requires that there be a contrastive relationship between the two functional units it coordinates. Her example illustrates this fact that but can only occur when the content of the upcoming unit(s) is in contrast with the content of the prior ones:

If Jews faced tolerance,

then I would not be against intermarriage.

but Jews face intolerance

but Jews provide tolerance.

We were kind to all,

So Jews provide tolerance,

but Jews face intolerance. (p.155).

Schiffrin further states that “Although the contrastive meaning of but gives it a narrower range of uses than and, its range is still fairly wide for the simple reason that contrastive relationships themselves are tremendously variable. Some times such relationships are transparent enough to be found in the semantic content of propositions, but others are buried within speakers’ and hearers’ culturally based world knowledge, or implicit in their expectations about each other and each other’s conduct.” (p.153). As mentioned previously, the purpose of this study is to examine whether non-native speakers of English know how to use CDMs appropriately to link S2 and S1 and make the sequences coherent.

Contrastive Discourse Markers

Fraser (1997), who coined the term “contrastive” to refer to this particular type of DMs, gives the following examples of CDMs:

(al)though, all the same , alternately, be that as it may, but, contrary to expectations, conversely, despite (doing) this/that, even so, however, in comparison (with/to this/that), in contrast (with/to this/that), in spite of (doing) this/that, instead (of (doing) that/that), nevertheless, nonetheless, on the contrary, on the other hand, rather (than (doing) this/that, still, whereas (3)

Fraser (1997) argues that there are specific “meaning distinctions” between each of the above mentioned CDMs. He states that these CDMs can be grouped together in terms of their co-occurence and divided into classes and subclasses in terms of placement restrictions. The following examples from Fraser (1997) illustrate clearly how CDMs im pose restrictions on the sequences they introduce:

(17) a) Fred is not a gentleman. On the contrary/*But, he is a rogue.

       b) A: Harry is quite tall. B: On the contrary/*But, he is really quite short.

       c) We didn’t leave late. But/*On the contrary, we arrived late.

       d) I don’t like this mess. But/*On the contrary, I understand how it occurred. (10)

As we can see from these examples, in (a) and (b) only the CDM on the contrary can make the two sequences coherent; using but instead of on the contrary does not seem logical. Similarly, in (c) and (d) but cannot be replaced by on the contrary. Fraser (1997) further points out that certain CDMs can co-occur, making the relationship between S2 and S1 meaningful. Consider this:

(8) a) The shipment of candy has arrived. But/However , don’t touch it.

     b) It’s alright Sue wasn’t here today. But/However, when will she be able to come ?

And,

(10) a) Take a letter. But/However , tell me if I am going too fast.

       b) Don’t smoke tobacco. Instead/Rather, chew the stuff. (6)

The above examples strengthen Fraser’s claim that certain combinations of CDMs can co-occur and certain combinations cannot. On the basis of their co-occurence and the

restrictions they impose on the segments they introduce to convey meaning, Fraser (1997) shows the relationships of CDMs in the following chart:

Chart 1 Relationship of CDMs (Fraser, 1997)

But

However

On the other hand Instead

In contrast Nevertheless Rather

In comparison

Conversely On the contrary

Fraser (1997) explains that the group of CDMs are divided into three different classes, based on their ‘core meaning’ (p.9). He argues that each of these CDMs imposes certain restrictions on the relationship between S2 and S1. To quote him,

The largest class, headed by but, imposes the least restrictions between S2 and S1 with which it is contrasted. The restrictions imposed by but are different from those imposed by instead/rather and on the contrary, such that where one of these classes can occur, the other two cannot. (9)

It is obvious from the above chart that but, however, and nevertheless; and instead and rather can co-occur, but but and instead; nevertheless and instead; and instead and on the contrary cannot. For the purposes of this study, it is assumed that native speakers of English intuitively know the distinction between each CDM and the restrictions Fraser’s classes and sub-classes of CDMs impose on meaningful sequences. This study will examine whether Russian speaking advanced ESL learners have internalized this knowledge, and to determine in what respect their use of CDMs differs from that of native speakers of English.

General Methods

Subjects

Ten native speakers of English who were taking an introductory course in linguistics at Boston University and ten advanced Russian ESL students were selected for this study. The native speakers of English were living in Boston and its surrounding areas. To this investigator’s knowledge, most of them were monolinguals and spoke English well.

The Russian ESL students were taking an advanced course in developmental reading and writing at Newbury College. Classes met Monday through Thursday from 9am to 12 p.m . They had 12 hours of intensive English per week and their instructor was a female native speaker of English. Their age varied from 35-40. At the time of the experiment, most of them were living in the Boston area.

Even though they were not fully competent in their second language, i.e. English, the Russian students were highly educated in their first language, and had had at least 6 years of English in Russia before they came to the US. Most of them had pursued professional careers while they lived in Russia. They started learning English in Russia at the age of 16. Their instructors were Russians, and the use of English, which mainly consisted of reading and writing, was restricted to the classroom. 

They reported that they seldom had an opportunity to practice conversational English in Russia. All of them spoke Russian in Russia to communicate with their parents, relatives, and friends. This was contrary to the situation in America where they had to use English for academic and communicative purposes. However, they frequently used Russian among themselves and with their program coordinator at Newbury College.

Data Collection/Analysis Procedures

Cloze Test

A cloze test, consisting of 30 multiple -choice questions, was given to both groups, i.e. native English speakers and Russian ESL students, at different times (see Appendix A). The 30 questions included but, however, nevertheless, despite this/that, in contrast, on the other hand, on the contrary, and instead. The subjects, both native and non-native speakers of English, were presented with two sequences of sentences in each question. Three CDMs were listed under each sequence in the space between the sentences. For each CDM, they had to determine whether it would be acceptable as a link between the two sequences. They had three choices to answer each question: put a “+” in the space before the connective if they were sure a particular CDM could link the two sequences; a “-” if they were sure the CDM could not link the two sequences; and a “? ” if they were unsure whether or not it could connect the two. Sometimes all three CDMs could link the two sequences of sentences, some times all three CDMs could not link the two sentences, and sometimes some of them could link the two sentences and some of them could not in a particular question.

This investigator discussed the possible answers to all 30 questions with a professor of linguistics at Boston University who is a native speaker of English, heads the applied linguistics program , and has taught courses in linguistics for almost 30 years. After a lengthy discussion, the results were summarized in a table (see Appendix B).

The native speakers finished the test in approximately 15 minutes. In contrast, the Russian students took almost a half hour to complete the test. This could be partly attributed to the fact that they were not allowed to use their Russian-English or English- English dictionaries during the test. The researcher assumed that since the Russian students were advanced learners of English as a Second Language, they would know most of the vocabulary words on the test.

Results

A close scrutiny of the native speakers’ test answers (see Appendix C) and the Russian students’ test answers (see Appendix D) revealed a great deal of difference in term s of what they thought were linkable and unlinkable CDMs. First, I shall discuss the native speakers’ test answers and then analyze the answers given by the Russian learners of English. Then I shall discuss the possible reasons for the correct and incorrect choices m a de by the learners. For those CDMs which can connect S1 and S2 in a particular context I shall use the term ‘linkable’. Similarly, I shall use the term ‘unlinkable’ to refer to those which cannot link the two sequences of sentences in a given situation.

Native Speakers’ Test Answers

But

As shown in table 1, there were 12 possible occurrences of but , and there were three instances where but could not have occurred. 5 out of 10 native speakers agreed that but could link the two sentences 12 times. 3 of them chose 11 possible occurrences. 1 speaker used it 9 times. In other words, in contrast to the others, this speaker thought that but could not have occurred on at least three occasions. Interestingly, 1 speaker thought but was acceptable 13 times, i.e.,

Table 1 - Native Speakers’ Test Answers

Discourse M a r k e r	C h o i c e	1	2	3	4	5	6	7	8	9	1 0
But	+ 12 - 3	9 2	11 2	12 2	11 3	12 3	12 3	11 4	12 3	12 3	13 1
However	+ 9 - 4	7 3	7 5	8 4	8 4	7 6	8 5	7 3	7 5	10 2	10 2
Nevertheless	+ 5 - 6	6 5	7 3	5 6	5 6	6 5	5 5	7 4	5 6	6 5	6 4
Despite this/that	+ 6 - 4	6 3	6 4	6 4	6 4	5 5	5 5	6 4	5 5	6 4	6 4
In contrast	+ 3 - 4	3 3	4 3	3 4	3 4	2 4	3 4	3 4	3 4	3 4	4 3
On the other hand	+ 3 - 10	3 10	2 11	3 10	2 11	2 11	2 10	3 10	2 11	3 10	4 8
On the contrary	+ 3 - 6	4 5	3 5	3 6	3 6	3 6	2 7	2 7	3 6	3 6	3 6
Instead	+ 5 - 7	5 5	5 7	5 7	5 7	6 6	5 6	5 7	5 7	4 8	4 7

in one question where but could not have possibly linked the two sequences of sentences, one native speaker thought it could.

On the other hand, there seemed to be a consensus among the speakers as to the restrictive use of but, i.e., most of them agreed almost unanimously where but couldn’t possibly link the two sequences of sentences. Barring one speaker, who thought this particular CDM was unlinkable only once, most of them agreed that it could not have linked the two sequences of sentences in a given question 3 times. There were, however, 3 speakers who chose but as non-occurring twice. Interestingly, as evident in the table, even some native speakers were unsure, at least once, whether or not it could connect the two sentences.

However

This CDM can almost always co-occur with but. Perhaps because of its coexisting nature with but, the speakers’ responses showed a s light variation. For example, most speakers chose the possible occurrence of however 7 or 8 times out of a possible 9, but two speakers thought it could occur 10 times. There were 4 instances where however could not have occurred. The speakers’ possible choices varied from 6 to 5 and 4 to 3. Two speakers were not sure about its linkability on one occasion.

Nevertheless

Once again, since nevertheless can co-occur with but and however, the speakers’ responses ranged from 7 to 5. It should be mentioned that there were 5 instances where the use of nevertheless was possible. On the other hand, seven speakers’ choice of non- occurrence fell between 6 and 5. Two of them thought it could not occur 4 times, and 1 speaker chose it 3 times. Interestingly, two speakers were unsure whether it could make the two sequences logical. Table 1 shows that nevertheless could not have occurred 6 times.

Despite this/that

Seven speakers unanimously chose this CDM to connect the two sequences 6 times, which was the possible number of occurrences. However, three of them thought that it could not have linked the two sentences in the given context; their response was 5. As for the non-occurrence of this particular CDM, their response was mixed. While 6 speakers correctly answered 4, 3 of them decided it could not occur linking S2 and S1 5 times. In other words, these speakers thought it could not occur where it should have occurred once, since there were 10 instances where despite this/that was one of the choices. 1 speaker was not sure about its use once.

In contrast

As regards the possible occurrence of this CDM, there seem s to be an agreement among most of the native speakers. 7 speaker s correctly chose it as a link between S2 and S1 4 times, whereas 2 of them thought it could occur 4 times, and 1 speaker chose it only twice out of a possible 4. This speaker was unsure whether in contrast could connect the two sentences on one occasion. In term s of the non-occurrence of this CDM in certain cases, almost everyone agreed that it could not link S2 and S1 4 times; only two speakers answered no 3 times. As with other CDMs, two speakers were not sure about its use once.

On the other hand

The number of occurrences where on the other hand could have occurred was 3. Only 4 speakers guessed it 3 times as a possible choice. Surprisingly, 5 speakers chose it only twice to link the two sequences of sentences in order to make them logical, meaning they didn’t’ think the use of this CDM was appropriate on one occasion. 1 speaker selected it 4 times as a possible connective. On the other hand, their responses to the non-occurring on the other hand ranged from 10 to 11 for 9 speakers. Only one speaker thought it could not occur 8 times. There were two speakers who were unsure about its use once.

On the contrary

As it had happened before, while 7 speakers correctly chose on the contrary as a possible link between S2 and S1 out of 3 possible instances, 2 thought it could connect the two sequences only twice, and 1 chose this CDM 4 times. In other words, while 2 of them thought it could not occur where it should have occurred, 1 speaker thought it could occur where it should not have occurred. As regards the non-occurrence of this CDM, their choices were in the neighboring area. As shown in table 1, on the contrary could not link the two sequences of sentences 6 times. Most speakers’ responses fell between 6 and 5, whereas two of them thought it could not link 7 times. Also, 1 speaker was unsure whether or not on the contrary could function as a possible link between S2 and S1 on one occasion.

Instead

Since the use of instead is restricted and presupposes a negative meaning in S1, there was not much variation among the choices given by the native speakers. As shown in table 1, the possibility of this CDM occurring between the two sequences was 5. Most speakers agreed that it could connect S2 and S1 5 times. Two of them considered it a possible choice 4 times, which means they didn’t think it was linkable once. On the other hand, six speakers guessed that it could not occur between the two sentences 7 times, which was the correct possible choice. While two thought this CDM couldn’t occur 6 times, one chose it 8 times as a non-occurring CDM, and another one selected it 5 times. Two speakers were confused about its use once and one twice.

As the data show, there is a slight variation in the judgment of the native speakers - both in terms of the possible occurrences of CDMs and the restrictions that are imposed on them by their core meanings - which could be attributed to a gamut of factors. Certain social and linguistic variables such as age, class, status, language background, and education may influence their judgment. This is not alarming because sometimes the distinctions among CDMs may not be as precise as we propose. Native speakers tend to have idiosyncratic ways of judging what is appropriate in their language and what is not.

Furthermore, since it is not the focus of this study to examine the acquisition of English CDMs by native speakers of English, and since the differences in their judgment are not large enough to draw serious attention, it is reasonable to ignore them in or variation in the native speakers’ choice of CDMs for the purposes of this study.

Advanced Russian ESL Students’ Test Answers

As shown in table 2, there was a great deal of variation among the Russian speaking advanced ESL students’ test answers. Numbers 1-10 represent the subjects who took the test, and the numbers in the “Choice” column are possible answers; “+” means linkable CDMs in a specific context, and “-” denotes unlinkable CDMs in the same context.

Table 2 - Advanced Russian ESL students’ Test Answers

Discourse M a r k e r	C h o i c e	1	2	3	4	5	6	7	8	9	1 0
But	+ 12 - 3	10 5	14 1	11 4	13 2	11 4	9 6	10 5	8 7	6 7	12 3
How e ver	+ 9 - 4	11 2	12 1	7 6	11 2	8 5	4 9	7 6	10 3	7 6	8 5
Nevertheless	+ 5 - 6	4 7	8 3	1 10	4 5	6 4	4 7	2 9	7 4	4 5	7 4
Despite this/that	+ 6 - 4	8 1	7 3	4 4	8 2	6 4	4 6	5 5	5 5	4 4	3 7
In contrast	+ 3 - 4	3 4	7 0	5 2	5 1	5 2	4 3	1 6	3 4	5 1	4 3
On the other hand	+ 3 - 10	4 9	4 9	6 7	6 6	4 8	4 9	5 8	2 11	3 10	11 2
On the contrary	+ 3 - 6	4 4	4 5	3 6	3 2	4 5	2 7	1 8	2 7	2 5	2 7
Instead	+ 5 - 7	5 5	4 8	4 8	7 5	5 7	4 8	4 8	4 8	4 7	5 7

For question 1, all subjects, except S1, S3, and S10, judged correctly that on the other hand could not link the two sequences of sentences in this particular context (see Appendix D). Furthermore, S1, S 3, and S6 incorrectly judged that on the contrary was a linkable CDM. While S1 and S3 were unsure as to the linkability of despite that, 8 subjects made the correct guess; their answers matched with the possible choice.

Compared to question 1, more subjects had difficulty answering question 2. S2, S5, S8, S9, and S10 correctly judged that nevertheless could occur between the two sequences of sentences, whereas S1, S3, S6, and S7 thought it couldn’t link the two sequences; S4 was not sure whether this particular CDM could link S2 and S1. Only 4 subjects correctly guessed that despite that couldn’t occur between S2 and S1, but S1, S2, S4, and S5 thought this CDM could link the two sequences of sentences. S3 and S9 were unsure about its use in this particular example. All the subjects judged correctly that besides nevertheless, however was also a possible connective.

In answering question 3, 5 subjects (S1, S3, S6, S8, and S10) m a de a m i stake by choosing on the other hand as a linkable CDM. Only 4 subjects made the correct guess. S5 was unsure as to the use of on the other hand in this example. On the contrary, however, didn’t seem to be problematic to most of the subjects. While seven subjects thought it was linkable, only three subjects judged that it was unlinkable. Interestingly, six subjects thought that in contrast was linkable where it couldn’t have occurred between the two sequences of sentences. Only S6, S7, S8, and S10 correctly judged that in contrast couldn’t link S2 and S1 in this context.

Except for S1, who thought that but couldn’t be used as a connective in question 4, all the subjects made the right guess. However, most subjects who had seemingly no difficulty with but incorrectly guessed that however could link S2 and S1. It m a y be the case that they sim p ly ignored the ‘though’ in S2, “Mary seem s all right, though.” Only S6, S7, and S9 answered correctly, choosing however as unlinkable. While 6 subjects correctly judged that nevertheless couldn’t make the contrastive relationship between the two sequences of sentences meaningful, 4 subjects thought this CDM was linkable in this context.

The subjects seemed to have an easier time judging the use of but in question 5. Except S7 and S9, all the subjects correctly judged that but was linkable. Similarly, except S6, S7, and S10, the rest of the subjects guessed that however could link the two sequences of sentences. Compared to but and however , on the other hand was a bit difficult for them . While 5 subjects chose it as unlinkable, which matched the correct answer, 5 subjects thought it was linkable.

In answering question 6, six subjects correctly guessed that nevertheless could occur between the two sequences of sentences. S4, S6, and S7 judged that it was unlinkable in this particular situation. S9 was unsure as to the use of this CDM. Judging the linkability of despite that in this question was relatively easier f o r the subjects. Except S10, all the subjects answered correctly. Similarly, almost all the subjects, except S1 who was unsure as to the use of this CDM, showed good knowledge of the use of instead in this context. They all thought that instead couldn’t occur between S2 and S1, which was the correct answer. It should be mentioned that as compared to but, however, and nevertheless, the use of instead is much more restricted and presupposes an explicit negative meaning in S1 (Fraser, 1997).

Interestingly, only six subjects made the right guess as to whether instead was linkable or unlinkable in question 7. S4, S6, and S10 thought it could link the two sequences of sentences, and S1 was unsure a bout its use. The second choice in this question, which was however, was easy for them. A total of nine subjects correctly judged that however could possibly occur between S2 and S1. Only S10 thought it otherwise. It seems that a majority of the subjects had not learned the proper use of nevertheless. While S2, S5, S8, and S10 correctly judged that it could make the relationship between S2 and S1 meaningful, five subjects thought it couldn’t occur between the two sequences of sentences. S9 wasn’t sure whether nevertheless could be used as a connective in this case.

As mentioned previously, the subjects seemed to know the use of but as a CDM well. Barring the exception of S8 and S9, all the subjects judged that but was a possible link between the two sequences in question 8. Also, they correctly judged that however could be used as a connective in this context. However, S10, who had correctly judged the linkability of but, thought that however couldn’t link the two sequences. Instead didn’t pose a problem at all. It seem s as if the subjects knew the fact that the use of instead requires a negative meaning in S1.

The first choice in question 9 was on the other hand. The function of this CDM as a connective is more specific than however (Fraser, 1997). Perhaps this is why many subjects had difficulty deciding which one was linkable and which was unlinkable. Except S7 and S8, who correctly judged that on the other hand couldn’t link S2 and S1, 8 subjects thought it was linkable. Deciding whether or not despite that could be used as a connective in this question, 8 subjects correctly answered that it couldn’t link the two sequences of sentences. Only S1 and S2 thought it was linkable, which didn’t m a tch the possible answer. All the subjects, with the exception of S1, correctly answered that in contrast was the only possible connective.

For question 10, the subjects had to choose from but, instead, and nevertheless. While seven subjects answered correctly that but couldn’t link the two sequences, S1, S2, and S4 misjudged the use of this CDM. Similarly, 8 subjects chose instead as a possible connective in this particular context. Only S1 and S10 thought instead couldn’t link S2 and S1. It seems that S1 was not familiar with this particular CDM, because he was the only subject who thought nevertheless could make the relationship between S2 and S1 meaningful. Everyone else judged that nevertheless was not a possible choice in this case.

Most subjects seemed to have little difficulty in answering question 11. Eight subjects, including S1, correctly chose in contrast as an unlinkable connective, whereas S2 and S4 thought it could link the two sequences of sentences. Similarly, most of them judged that nevertheless could connect S2 and S1 in a meaningful way. Only S3, S4, and S7 thought it couldn’t connect the two sequences. Barring the exception of S9, all the subjects correctly judged that despite that could occur between S2 and S1.

On the other hand, but, and in contrast were the three choices for question 12. Unlike question 9, where the subjects misjudged the use of on the other hand, 8 subjects correctly judged this time that this CDM couldn’t possibly link the two sequences; only S4 and S10 thought it could occur between the two. As mentioned previously, the subjects seemed to know the proper use of but as a CDM. All of them answered correctly that but could be used as a possible connective in this case. In the case of in contrast, eight subjects judged that it was unlinkable. S2 thought it was linkable, and S9 was unsure about its use.

Answering question 13, seven subjects rightly guessed that but was a possible connective between the two sequences. However, S3, S8, and S9 thought that it couldn’t occur between S2 and S1. W ith the exception of S1, nine subjects correctly judged that instead couldn’t connect the two sequences in this context. In contrast also didn’t seem to be problematic to the subjects, because eight of them correctly judged that it could be used as a connective; only S1 and S7 thought that this CDM was unlinkable.

In question 14, the subjects had to choose from nevertheless, despite that, and on the contrary. As mentioned previously, nevertheless seemed to be problematic to most subjects. While six subjects guessed that nevertheless couldn’t link S2 and S1, which matched the possible answer, S6 and S7 thought it could be used as a connective; S4 and S5 weren’t sure about its use. Only S1 and S4 misjudged the use of despite that. Eight subjects answered correctly that this CDM could not occur linking the two sequences of sentences. In term s of using on the contrary as a connective, six subjects correctly judged that it could occur between S2 and S 1. S6, S7, and S8 thought it couldn’t link the two sequences, and S4 wasn’t sure about its use.

Most subjects did fairly well on question 15. Only S2 and S4 misjudged the use of nevertheless in this particular context. Other than that, the subjects’ answers were on target.

Six subjects correctly answered that but was linkable in question 16. S1, S5, S6, and S7 thought it couldn’t possibly occur linking the two sequences of sentences. However also seemed to be relatively easier for the subjects. Only S5 and S6 misjudged its use. Seven subjects correctly chose despite this as a linkable CDM. S5, S6, and S10 incorrectly chose it as unlinkable.

The first choice in question 17, but , confused half of the subjects. While five subjects correctly judged it as a possible connective, five subjects thought it couldn’t possibly link S2 and S1. However also seemed to be confusing to them . With the exception of S4, nine subjects chose it incorrectly as an unlinkable CDM. However, they didn’t have much difficulty with in contrast. Eight subjects correctly judged that this CDM could link the two sequences; S7 didn’t think it could link the two; and S4 was unsure about its use.

Most subjects did fairly well on question 18. Eight subjects answered correctly that but was a possible connective; S1 answered incorrectly; and S9 was unsure. Similarly, eight subjects correctly judged the use of however. Only S3 and S7 thought it was not a possible choice in the given context. While eight subjects had no difficulty ruling out on the contrary as a possible connective, S1 and S4 were not sure whether this CDM could connect the two sequences of sentences or not.

In answering question 19, eight subjects correctly judged that on the other hand couldn’t possibly occur between the two sequences. S6 misjudged its use, and S4 was unsure about its use. Seven subjects rightly ruled out on the contrary as a possible connective. S1 thought it could be used as a connective, and S4 and S9 were unsure. All of the subjects answered correctly that instead was the only possible choice.

Question 20 also didn’t cause the subjects much problem. All of them, except S6, judged that but could occur linking the two sequences of sentences. As usual, with the exception of S1, all of them ruled out instead as a possible choice. However, on the other hand was a bit difficult for them . Six subjects incorrectly judged that it could link the two sequences. Only S1, S3, S6, and S8 answered correctly.

In a similar way, most subjects did well on question 21. Six subjects chose but as a possible connective. S6, S7, and S8 didn’t think it could connect the two sequences, and S9 was unsure about its use. Eight subjects correctly judged however as a possible choice; only S6 and S9 misjudged its use. Similarly, eight subjects answered correctly in terms of discarding on the other hand as a possible connective. S4 and S10 misjudged that it could occur between the two units.

But, instead, and on the contrary were the three choices for question 22. The subjects had no problem discarding the use of but in this context. Almost all of the subjects answered correctly that instead was the only possible choice in this case; for some reason, S6 thought it couldn’t make the relationship between S2 and S1 meaningful. In contrast, only f our subjects correctly judged that on the contrary was not a possible connective. While S1, S2, S5, and S8 thought it could link the two sequences, S4 and S9 were unsure whether it was a possible choice.

The subjects’ response to the use of nevertheless in question 23 was mixed. Five subjects judged that nevertheless was a possible connective, and five subjects considered it unlinkable. All the subjects agreed that despite this was a possible choice. Similarly, barring the exception of S8, all of them chose but as a possible connective between the two sequences.

In answering question 24, eight subjects correctly ruled out on the other hand as a possible connective. Only S7 and S10 thought it could link the two sequences. On the contrary seemed to be a bit confusing to them. While four subjects correctly chose it as a possible link between S2 and S1, six subject s judged that it couldn’t occur linking the two sequences. Also, however was correctly chosen as unlinkable by seven subjects, but S1, S2, and S10 thought it could link the two sequences.

Question 25 seemed to be easier for the subjects, because most of them answered it correctly. Eight subjects discarded on the other hand as a possible link. Only S7 and S10 chose it as a possible connective. Nine subjects ruled out on the contrary as a possible link between the two sequences. For some reason, S4 judged that it could occur linking S2 and S1. While six subjects correctly judged but as a possible connective, four subjects (S4, S6, S8, and S9) didn’t think it could occur between S2 and S1.

The subjects had to choose from on the other hand, on the contrary, and nevertheless for question 26. Seven subjects judged that on the other hand was not a possible link, whereas S3, S4, and S10 thought it could occur linking the two sequences of sentences. Also, nine subjects ruled out on the contrary as a possible choice for this question. Only S6 judged that it could be used as a link between the two sequences. Nine subjects chose nevertheless as a possible connective, which matched the correct answer, whereas S3 misjudged its use. Ignoring the few mistakes the subjects made, it seems that they knew the possible choices in this particular context.

Question 27 also didn’t seem problematic to the subjects at all. All ten subjects judged that but was the only possible connective. Nine subjects ruled out instead as a possible link between S2 and S1; only S4 judged that it could link the two sequences. Similarly, eight subjects discarded nevertheless as a possible connective. S4 and S10 misjudged its use.

The subjects’ answers to question 28 were almost on target. Nine subjects, with the exception of S1, chose instead as the only possible connective. All of them thought that on the other hand was not a possible choice, and seven subjects judged that however was also not a possible link between the two sequences of sentences. In contrast, S1, S2, and S10 thought that however could occur linking the two sequences.

The subjects’ answers to question 29 were slightly off base. While seven subjects correctly discarded instead as a possible connective, S4 and S10 chose it as a link between the two sequences; S9 was unsure about its use. Similarly, seven subjects rightly judged that however was the only CDM that could possibly link the two sequences of sentences in this context, whereas S3, S6, and S10 thought that it couldn’t occur between the two. In contrast, however, caused most of the subjects some problem in judging its correct use. Eight subject s thought it was a possible connective, which didn’t match with the possible answer. Only S7 and S8 correctly judged that it wasn’t a possible choice.

The last question received a mixed response from the subjects. While eight subjects correctly judged that instead was not a possible connective, and S1 and S5 thought it was, five subjects thought however could occur linking the two sequences of sentences, and five subjects thought it couldn’t. The subjects also made many mistakes in judging the linkability of despite this. Only S1, S4, and S5 rightly judged that despite this was a possible choice for this question; seven subjects misjudged its use by choosing it as unlinkable.

A few significant points emerge from these observations. If we examine the range of uses for particular markers, we find that some Russian subjects have a wider range of uses for certain markers. In the da ta, it seemed to be generally the case that some subjects used markers appropriately in a range of functions, whereas some used markers with a m o re lim ited range of functions. Some of them didn’t know how to use certain markers in certain contexts.

These differences in the subjects’ performances can be attributed to a gam u t of factors such as proficiency levels, exposure to the target language, language transfer, etc. Perhaps a sub-category of proficiency is lexicon. It should be noted that the subjects were not allowed to use their dictionaries, because it was assumed that since they were advanced learners of English, they would be familiar with the vocabulary words on the test. It is likely that the subjects didn’t know certain vocabulary words and chose incorrect answers. Furthermore, there may be a correlation between the degree of interaction with native speakers, besides classroom instruction, and the use of markers. Given the differences in their performances, it is reasonable to assume that interaction with native speakers of the target language facilitates the acquisition of those markers which are not available in the first language. Second language learners use information- processing strategies or problem -solving procedures, which make adult language learning quite different from child language acquisition. Although the input processing strategy may not work sometimes, “the insight that acquisition involves input-processing strategies of some kind is important and should be pursued” (White, 1991). This study didn’t pursue this aspect of SLA. The above-mentioned factors need to be examined empirically to draw substantial conclusions.

Narrowing of Research Questions

With these initial results then, we may begin to narrow down our research questions and hypotheses. First, a future large-scale study should ask what is the full range of markers available to L2 learners at different proficiency levels. Our limited data of course could not answer this question. A second part to this first question is what role does the L1 play in acquisition. That is, do learners more easily acquire those markers which have equivalents in the first language?

Second, a future large-scale study should ask what is the full range of functions of markers available to learners at different levels of proficiency. Related to this is the question of the role that the L1 plays. That is, do second language learners more easily acquire those functions of markers which have equivalents in the first language? Of course, all of these questions will lead to further more refined questions, but this investigator believes that at least these questions need answering at this early stage.

Refinement of Method

A practical question, which arises at this point, is what sorts of methods are appropriate for answering these research questions. This pilot study was flawed methodologically on several grounds. First, I tested subjects from only one proficiency level (determined roughly by the number of years spent in the United States). Second, only a few CDMs were selected for this study, and the subjects knew what the researcher was looking for. This could create some conscious or subconscious influence on their behavior. Therefore, data for future studies should involve distinct (carefully defined) proficiency levels, and a wider range of CDMs should be examined.

There are two distinct methods which will be appropriate to answering the research questions proposed earlier. First, a large-scale quantitative study is the only way to make adequate generalizations about groups of learners at different proficiency levels, etc. In order to generalize about any developmental sequence which occurs in the acquisition of discourse markers, a large amount of data must be collected from a large number of individuals. This might involve, say, at least three distinct proficiency levels, with ten members each, making a total of thirty subjects. The test should include a wider range of CDMs to elicit the full range of DMs available to the learners.

Second, a large-scale qualitative study is the only way to discover the range of uses of a single marker. This kind of study could make use of the same data as the quantitative study, but instead of counting the number of discourse markers correctly and incorrectly used, one would have to analyze the different functions to which particular markers are put. This could then be correlated with proficiency levels. Such studies could focus in detail on one or a few markers.

For both the quantitative and qualitative studies, one would also need a fairly comprehensive study of the CDMs belonging to both the first language and the target language. For the quantitative study, this would involve defining the set of markers in both the L1 and L2. For the qualitative study, this would involve defining the various functions of different markers in both the L1 and L2.

Directions for Future Studies

In sum, this investigator believes that SLA studies would benefit from an examination of the acquisition of DMs in general and CDMs in particular. At this incipient stage, there are three main questions, which should give direction to future studies. First, what is the full range of DMs available to L2 learners at different proficiency levels? Second, what is the full range of functions of markers available to learners at different levels of proficiency? And last but not the least, what factors determine and influence the acquisition of CDMs? The answers to these initial questions require distinct quantitative and qualitative methods, the results of which will supplement each other.

References

Bazanella, C. 1990. Phatic connectives as intonational cues in contemporary spoken Italian.      

       Journal of Pragmatics , 14(4):629-47.

Blakemore, D. 1987. Semantic constraints on relevance. Oxford: Blackwell.

Blakemore, D. 1992. Understanding utterances. Oxford: Blackwell.

Ellis, R. 1987. Understanding second language acquisition. Oxford: Oxford University

       Press.

Ellis, R. 1996. The study of second language acquisition . Oxford: Oxford University

       Press.

Erman, B. 1987. Pragmatic expressions in English: A study of ‘you know’, ‘ you see’

       and ‘I mean’ in face-to-face conversation . Almqvist & Wiksell.

Fraser, B. 1987. Pragmatic formatives. The Pragmatic Perspective, ed. by J. Verschueren & M.      

       Pertuccellli-Papi. Philadelphia: John Benjamins.

Fraser, B. 1988. Types of English discourse markers. Acta Linguistica Hungarica, 38(14):19-33.

Fraser, B. 1990. An approach to discourse markers. Journal of Pragmatics, 14:383-95.

Fraser, B. 1997. Contrastive discourse markers in English. Boston University Manuscript.

Halliday, M. and R. Hasan. 1976. Cohesion in English . London: Longman.

Holker, K. 1991. Franzosishch: partikelforshung. Lexikon der Romanistishchen Linguistik , Vol 

        V(1):77-88. Niemeyer.

Hovy, E. 1994. Parsimonious and profligate approaches to the question of discourse structure 

        relations . London: Longman.

Katriel, T. and Dascal M. 1977. Between semantics and pragmatics: the two types of but” - Hebrew 

       “aval” and “ela”. Theoretical Linguistics , 4:143-72.

Knott, A. and R. Dale. 1994. Using linguistic phenomena to motivate a set of coherence relations. 

        Discourse Processes , 18(1):35-62.

Ostman, J-O. 1989. On the language-interval interaction of prosody and pragmatic particles. Levels 

        of Linguistic Adaptation , ed. by J Verschueren. Philadelphia: John Benjamins.

Polanyi, L. and P. Scha. 1983. The syntax of discourse. Text 3:261-70.

Quirk, R., Greenbaum , S., Leech, G. and Swartvik, J. 1985. A comprehensive grammar of the 

        English Language . London: Longman.

Redeker, Gisela. 1991. Review article: Linguistic markers of discourse structure. Linguistics,

        29(6):1139-72.

Schiffrin, D. 1987. Discourse Markers. Cambridge University Press.

Schoroup. 1985. Common discourse particle in English conversation: like, well, y’know. Garland.

Stubbs, M. 1983. Discourse analysis. University of Chicago Press.

Van Dijk, T. 1979. Pragmatic connectives. Journal of Pragmatics, 3:447-56.

Vande Kopple, W. 1985. Some exploratory discourse on meta-discourse. College Composition and 

        Communication. 36.

White, L. 1991. Second language competence versus second language performance: In Eubank, L.

        (ed.). Point Counterpoint: Universal Grammar in the Second Language . Philadelphia: John 

        Benjamins.

APPENDIX A
Name _________ Age _____ 20-30 _____ 30-40 _____ 40 +
Level ___________ Nationality ____________ First language ______________

Read each of the following sequences of sentences carefully. Under each sequence in the space between the sentences, are listed three connectives. For each connective, determine if you think it would be acceptable as a link between the two sentences. That is, does it sound right to you? If you are sure it can occur linking the two sentences, put a “+” in the space before the connective. If you are sure it cannot occur linking the two sentences, put a “-” before the connective. If you are unsure, put a “? ”. Keep in mind that sometimes all three connectives may link the two sentences, and sometimes all three connectives may not link the two sentences, and sometimes some of them may link the two sentences and some of them may not.

Example:
Americans like to eat out. _____ the Chinese like to eat in.
_
+
_ But
_
+
_ However,
_
-
_On the contrary,

You would place a “+” before the “But” and “However” but a “-” before the “On the contrary”.

1. Mary wasn’t in love with John. _______ she married him .
___ On the other hand,
___ On the contrary,
___ Despite that

2. I can explain it orally. _______ if you insist, I will put it in writing.
___ Nevertheless,
___ Despite that,
___ However,

3. I hoped he would go early. __________ he stayed till midnight.
___ On the other hand,
___ On the contrary,
___ In contrast,

4. John doesn’t look very happy. _______ Mary seem s all right, though.
___ But
___ However,
___ Nevertheless,

5. You can phone the doctor if you like. _______ I doubt whether you will get him to com e out on a Saturday night. ___ But
___ However,
___ On the other hand,

6. He is poor. ________ he is satisfied with his situation.
___ Nevertheless,
___ Despite that,
___ Instead,

7. W e don’t have to go to the party. _______I will go.
___ Instead,
___ However,
___ Nevertheless,

8. We left late. _______ we arrived home on time.
___ But
___ However,
___ Instead,

9. John is very friendly. _______ his brother, Bob, is very rude.
___ On the other hand,
___ Despite that,
___ In contrast,

10. Harry didn’t go to the county fair. _______ he stayed home and did his homework.
___ But
___ Instead,
___ Nevertheless,

11. John felt sick. _______ he still went to school.
___ In contrast,
___ Nevertheless,
___ Despite that,

12. He’s good looking. _______ that isn’t going to get him a job in this market.
___ On the other hand,
___ But
___ In contrast,

13. John runs marathons. _______ Jim sits home and sleeps.
___ But
___ Instead,
___ In contrast,

14. Harry is no gentleman. _______ he is a rude bore.
___ Nevertheless,
___ Despite that,
___ On the contrary,

15. It m a y rain today. _______ it m a y not.
___ Nevertheless,
___ Despite that,
___ On the other hand,

16. John was late in leaving home. _______ he arrived on time.
___ But
___ However,
___ Despite this,

17. The Democrat’s mascot is a donkey. _______ the Republican’s is an elephant.
___ But
___ However,
___ In contrast,

18. The shipment of candy has arrived. _______ I don’t want you to touch it.
___ But
___ However,
___ On the contrary,

19. Don’t smoke tobacco. _______ chew this stuff.
___ On the other hand,
___ On the contrary,
___ Instead,

20. Keep the faith. _______ don’t be stupid about it.
___ But,
___ Instead,
___ On the other hand,

21. I know that sue wasn’t here today. _______ when will she be able to come ?
___ But
___ However,
___ On the other hand,

22. Don’t sit down. _______ I want you to take a walk with me .
___ But
___ Instead,
___ On the contrary,

23. John was late in leaving for the party. _______ he arrived on time.
___ Nevertheless,
___ Despite this,
___ But

24. Don’t stop. _______ keep going until you see the red house.
___ On the other hand,
___ On the contrary,
___ However,

25. Take a letter. _______ I don’t want you to send it right away.
___ On the other hand,
___ On the contrary,
___ But

26. I didn’t have enough money. _______ I took a vacation and went to Italy for two weeks. ___ On the other hand,
___ On the contrary,
___ Nevertheless,

27. I have been trying to explain it to you. _______ you are not listening to me .
___ But
___ Instead,
___ Nevertheless,

28. Don’t call your girlfriend. _______ call your mother. She will give you good advice.
___ Instead,
___ On the other hand,
___ However,

29. John can’t see very well. _______ he can hear perfectly.
___ Instead,
___ However,
___ In contrast,

30. He didn’t go to his father’s funeral. _______ he wasn’t sorry at all.
___ Instead,
___ However,
___ Despite this,