08 Artiga.qxp IBÉRICA 12 [2006]: 127-144 Abstract Epistemic lexical verbs (Hyland, 1998) constitute essential elements in conveying different interpersonal meanings, among those judgment, certainty, doubt, and evidence, aiding speakers both to project themselves into the speech and to maintain interesting relationships with audiences. Supporting Sinclair’s (1991) contention that meanings are clustered into lexicogrammatical patternings, and not in isolated items, this paper explores the phraseology accompanying the most representative epistemic lexical verbs (ELVs) in university lectures. By applying corpus-based methodology, we also illustrate the functional variability of think –the most salient of these lexical verbs–, especially the roles it performs when hedging the discourse and complying with politeness conventions as part of interactional and metadiscoursal strategies. At a textual level, results indicate that lecturing speech favors the proliferation of evaluation markers that appear to be arranged in several recurrent formal patterns. This clearly entails potential for pedagogical purposes and ESP syllabuses. Key words: academic speech, lectures, stance, lexical units, interpersonal pragmatics Resumen La interacción semántico-pragmática de la presencia del autor en la fraseología de las clases magistrales Los verbos epistémicos (Hyland, 1998) constituyen elementos esenciales en la transmisión de múltiples significados interpersonales, entre los que se incluyen el juicio, la certeza, la duda o la evidencia, contribuyendo de manera eficaz a que los hablantes se proyecten dentro del discurso y establezcan relaciones interesantes con la audiencia. De acuerdo con los postulados de Sinclair (1991) que argumenta que los significados no se codifican de forma aislada sino que tienden a formar estructuras lexicogramaticales, este estudio pretende explorar la fraseología típica de los verbos epistémicos (ELV) más frecuentes en clases magistrales. Asimismo, se ilustrará mediante la metodología del análisis de corpus, la variabilidad funcional de think, por ser el más frecuente de dichos verbos léxicos, en The semantic-pragmatic interface of authorial presence in academic lecturing phraseology María Rosario Artiga León Universidad de Zaragoza charomar21@wanadoo.es 127 08 Artiga.qxp 20/09/2006 13:51 PÆgina 127 particular las funciones que desarrolla matizando el discurso y conviniendo con los mandatos de la cortesía dentro de las estrategias de interacción y metadiscurso. A nivel textual los resultados indican que este tipo de discurso universitario favorece la proliferación de indicadores de evaluación que forman diversos patrones tipo, lo que implica un enorme potencial para su aplicación práctica en los currículos de Inglés para Fines Específicos. Palabras clave: discurso académico, clases de universidad, posición de autor, unidades léxicas, pragmática interpersonal Introduction Far from being neutral or straightforward towards the information conveyed writers/speakers tend to include affective meanings such as their own attitudes and evaluation in addition to propositional content. Over the last several years, linguists have become increasingly interested in the ways speakers and writers express judgments and even personal feelings to support their arguments and validate their claims. Such investigations have been carried out from different perspectives and under several labels, including “attitude” and “modality” (Halliday, 1994), “evaluation” (Hunston, 1994), “affect” (Ochs, 1989), “evidentiality” (Chafe & Nichols, 1986), “hedging” (Holmes, 1988; Hyland, 1996, 1998; Markkanen & Schröder, 1997), “appraisal” (Martin, 2000), and “stance” (Biber & Finegan, 1988, 1989; Beach & Anson, 1992; Barton, 1993; Conrad & Biber, 2000). Although these terms cover slightly overlapping areas, they all refer to writer/speaker point of view about the state of affairs or the information transmitted. This paper will primarily adhere to Biber and Finegan’s (1988) concept of stance, which refers to the ways speakers and writers convey their personal feelings and assessments in addition to propositional content. These authors define “epistemic stance” as the degree of certainty or evidence towards the content; it is roughly modalization in Halliday’s (1994) terminology, whereas “attitudinal stance” conveys more personal feelings, and “style stance” relates to the speaker’s or writer’s comment on the way in which the information in a clause is presented. Stance has been used as an umbrella word comprising several heterogeneous textual features. Among the devices deployed to index stance, epistemic lexical verbs (Hyland, 1998), heretofore ELVs, such as think, suggest, feel and seem are essential to encode judgment and carry out evaluative functions, both in written and spoken academic registers (Hyland, 1998; Biber et al., 1999). As stated above, these elements help perform the interpersonal function of language and are part of modality systems, which broadly refer to “a speaker’s attitude towards, or opinion about, the truth of a proposition expressed” and “extends to their attitude towards the situation or event described by a sentence” (Simpson, 1993: 47). MARÍA ROSARIO ARTIGA LEÓN IBÉRICA 12 [2006]: 127-144128 08 Artiga.qxp 20/09/2006 13:51 PÆgina 128 Whereas the proliferation of these stance markers in conversational settings has already been pointed out by several researchers (Biber et al., 1999; Hunston & Thompson, 2000), the situation in the academic spoken registers has been somewhat neglected so far and studies have been more sporadic (Mauranen, 2000, 2001; Poos & Simpson, 2002; Swales & Burke, 2003; Simpson, 2004). In lecturing, being this an instructional genre, speakers easily resort to ELVs by adopting various positions to project themselves into the speech: assessing knowledge, doubting some aspects, commenting on propositions, indicating the source of the information, agreeing/disagreeing, and making alliances and commitments, among others: We also know that academic speech tends to be heavily hedged, although, as Poos and Simpson (2002) have suggested, not only for the traditional reasons of modesty and uncertainty, but also (at least in instructive speech) as a way of socializing students into the sensitivities of a particular discipline. (Swales & Burke, 2003: 1) Information in lectures, although planned beforehand and possibly somehow previously rehearsed, is delivered online and interaction tends to share the characteristics of both face to face and spontaneous exchanges (Swales, 1990; Mauranen, 2000). Teachers generally favor audience oriented devices and appear accessible and encouraging providing personal support and allowing relaxed, informal encounters. From a different perspective, epistemic elements are not usually selected in isolation; rather, they tend to cluster in formulaic lexical chunks (Sinclair, 1991; Hunston & Sinclair, 2000). As Nattinger and DeCarrico (1992: 11) put it, they make “lexical phrases” typically “consisting of a syntactic frame that contains slots of various fillers, and run the gamut from completely fixed unvarying phrases to phrases that are highly variable.” These lexicalized composites seem to be fixed not only in their form but also in their functional application. Therefore, in the expression of meanings, it could be reasonable to describe sense and structure together, as Sinclair (1991: 65) concurs, “it seems that there is a strong tendency for sense and syntax to be associated.” He further argues that “speakers have meanings which they want to make and that these meanings naturally attract to themselves phraseologies which incorporate lexis and grammar” (Sinclair, 1991: 31). It is then likely that certain structural patterns select similar words for particular meanings. On pedagogical grounds, this suggests that rather than teaching isolated lexical elements, English for specific and academic purposes instruction should pay closer attention to complex patternings to convey meanings successfully since “fluent and idiomatic control of a language rests to a considerable extent on knowledge of a body of ‘sentence stems’ which are ‘institutionalized’ or ‘lexicalized’” (Pawley & Syder, 1983: 191) in order to help learners’ production achieve native-like accuracy. From a pedagogical point of view, in an attempt to bridge the gap between research THE SEMANTIC-PRAGMATIC INTERFACE IBÉRICA 12 [2006]: 127-144 129 08 Artiga.qxp 20/09/2006 13:51 PÆgina 129 and teaching, our study will be relevant to detect lecture comprehension verbs selecting recurrent lexical gambits, which will be of interest for those wanting to communicate effectively in the academic world. In the light of these considerations, my aim in this paper is twofold, firstly to identify the routine composites and to observe the frequency and recurrence of these clusters describing the correlates that accompany ELVs as pillars on which the conceptual load of lectures is framed. Secondly, I will focus on the verb think, to illustrate the pragmatic effects of ELVs within a context of use. Although any attempt to discuss and assess the implications this research might have for teaching English for Academic Purposes (EAP) and English for Specific Purposes (ESP) falls outside the scope and length of this article, some general pedagogical considerations will also be sketched out at the end of this paper. Analytical procedures The corpus chosen for the present study consists of 14 lectures, a total number of 152,810 words, chosen randomly out of the 62 lectures available in MICASE (Michigan Corpus of Academic Speech in English), and it is the only electronic compilation that contains samples of real academic speech events that I know of. MICASE is an appropriate resource since it includes authentic data recorded in a university setting and also covers four different domains of knowledge: Social Sciences and Education, Humanities and Art, Physical Sciences and Engineering, and Biological and Health Science. The lectures were analyzed by using the concordance program WordSmith Tools 3.0 (Scott, 1999), which allows rigorous contextual investigations and, additionally, enables horizontal and vertical study of concordances, essential in the identification of frequently-occurring sequences of items and phraseology. Corpus-based research was used to provide empirical evidence in order to reveal much more about grammatical features and to enable investigation of the actual patterns of use in addition to structural descriptions. In a preliminary stage of the research, elements not found to have epistemic meanings in the context of use were discarded, which happened in nearly half of the tokens (734 out of 1,347). We resorted to an expert native scholar (see acknowledgements) for clarification on some of the doubtful meanings these ELVs conveyed as this task was not always straightforward considering there exist fuzzy and overlapping instances difficult to discern in this slippery category of lexical verbs. Then, in the first stage of this study, a quantitative analysis of the frequencies observed by the different items under study was carried out to find out the most recurrent ELVs in the genre. Later, by using corpus-based analytical techniques, we MARÍA ROSARIO ARTIGA LEÓN IBÉRICA 12 [2006]: 127-144130 08 Artiga.qxp 20/09/2006 13:51 PÆgina 130 were able to document the different structural patterns by which speakers mark stance using several lexical and syntactic resources available. Finally, the focus of analysis was turned to the pragmatic domain to study the communicative functions the verb think carries out in the context of university lectures. Results and discussions Quantitative results In an initial search, the frequency distribution these ELVs score in the corpus was observed to identify which tokens are favored in university lectures. Following Hyland (1998), Biber et al. (1999), and Hunston and Thomson (2000), on the most recurrent lexical verb items conveying epistemic meanings to encode stance in speech and academic settings, the following verbs were selected: admit, appear, argue, assume, believe, calculate, claim, conclude, consider, doubt, estimate, exhibit, feel, figure, guess, imagine, indicate, infer, judge, not know, note, notice, predict, seem, show, speculate, suggest, suppose, suspect, and think. Results of the frequencies found in the corpus are illustrated in Table 1. THE SEMANTIC-PRAGMATIC INTERFACE IBÉRICA 12 [2006]: 127-144 131 LEXICAL VERB Total number of counts Counts with epistemic meaning Average number of counts per 10,000 words Comparative percentage admit 2 1 0.06 0.13% appear 36 10 0.65 1.36% argue 18 11 0.71 1.49% assume 41 35 2.29 4.76% believe 29 22 1.43 2.99% calculate 6 2 0.13 0.27% claim 8 3 0.19 0.40% conclude 9 6 0.39 0.81% consider 21 6 0.39 0.81% doubt 1 - - - estimate 23 1 0.06 0.13% exhibit 5 1 0.06 0.13% feel 58 15 0.98 2.04% figure 28 3 0.19 0.40% guess 25 15 0.98 2.04% imagine 23 6 0.39 0.81% indicate 6 3 0.19 0.40% infer 3 1 0.06 0.13% judge 10 3 0.19 0.40% not know 105 73 4.77 9.94% note 47 4 0.26 0.54% notice 45 12 0.78 1.63% predict 11 6 0.39 0.81% seem 57 57 3.73 7.76% show 90 10 0.65 1.36% speculate - - - - suggest 35 13 0.85 1.77% suppose 65 38 2.48 5.17% suspect 2 2 0.13 0.27% think 538 375 24.50 51.08% TOTAL 1,347 734 (55%) 48.03 100% Table 1. Total number of tokens of epistemic lexical verbs in lectures MICASE (152,810 words). 08 Artiga.qxp 20/09/2006 13:51 PÆgina 131 Quantitatively speaking, think scores the highest frequencies, more than 51% of the total number of ELV tokens, followed by other items such as not know (9.94%), seem (7.76%), suppose (5.17%), assume (4.76%), and believe (2.99%). In the case of know, only negative tokens expressing uncertainty were selected since no epistemic meanings were found in the affirmative form of this verb. All these six recurrent items at the top of the rank are favored by conversational settings (Biber et al., 1999), whereas other verbs which show insignificant or null frequencies (admit, calculate, claim, doubt, exhibit, indicate, infer, judge, speculate, and suspect) correspond to verbs more characteristic of written registers. Thus, our results corroborate previous findings1 on the subject by Hyland (1998) and Biber et al. (1999). In the following stage of this research, admitting that some associations in the dichotomy of form and function can be found in ELVs, we set to identify the lexical bundles that constitute specific linguistic realizations to mark stance. Phraseological correlates of epistemic lexical verbs in lectures Adhering to Sinclair’s (1991) and Hunston and Thompson’s (2000) contention that lexis and grammar constitute a single phenomenon and not two separate systems, this section illustrates how the language of stance tends to be realized in linguistic units displaying a varied range of structural possibilities. The analysis below will be centered on the six most frequent ELVs in the selected corpus, think, not know, seem, suppose, assume and believe (81.74% of the total number of counts) since the comparative frequencies of some of the remaining verbs were not considered significant. The corpus displays several grammatical and lexical patterns containing epistemic lexical verbs to convey different evaluation meanings as illustrated in Table 2: MARÍA ROSARIO ARTIGA LEÓN IBÉRICA 12 [2006]: 127-144132 ELVs Total tokens Passive ELVs Pattern 1: + that-clause Pattern 2: Parenthetical sentence Pattern 3: + to infinitive Pattern 4: + to epist. verb think 375 4 280 66 - 30 not know 73 - 5 5 - - seem 57 - 6 2 29 - suppose 38 34 2 1 33 - assume 35 2 19 2 - 4 believe 22 1 9 3 - 3 TOTAL 600 41 6.83% 321 53.5% 79 13.16% 62 10.33% 37 6.16% Table 2. Phraseological correlates of the most frequent epistemic lexical verbs in lectures MICASE (152,810 words). 08 Artiga.qxp 20/09/2006 13:51 PÆgina 132 Pattern 1: Personal/impersonal subject + epistemic lexical verb + (that)-clausal complement This pattern is by far the most recurrent verbal composite since all these epistemic markers allow this structure. In our corpus more than half of the frequencies 383 out of a total of 734 (321 out of 600, a 53.5%, in the most frequent verbs), show this patterning. Moreover, some authors even include these ELVs in the category of ‘believe-type verbs’ (Givón, 1993; Nöel, 2002). Although explicit that is frequent as illustrated in examples 1 and 2, that is more usually omitted (example 3). In these three instances the matrix verb conveys the speaker’s stance towards the content of the forthcoming proposition; in the first example the lecturer indexes the source of the evidence for his/her claim whereas in the second and third ELVs are used to somehow add speculation towards the content of the propositions. (1) […] but they still don't metastasize as well as D-one-twenty-two. fo- so this suggests that there are additional factors that the immune system yes is one factor but there are other factors […] (LEL175SU106)2 (2) […] so the world that Marx was seeing, was one of increased worker exploitation, and Marx believed that as this worker exploitation became utterly unbearable […] (LEL220JUO71) (3) […] so can i explain for example the fact that if i go, how's everybody doing? you all assume i'm, more likely happy, than unhappy […] (LEL500JU034) Apart from a that-clause, other alternative subordinate sentences can be found as is the case of why, where, what, whether and if, these being commonly associated with not know. Thus, in example 4 not know hedges a subordinate sentence with what mitigated by a proliferation of other stance markers (‘I think’, ‘might’, ‘exactly’, conditional sentences …), and in example 5 the verb conveys uncertainty towards the following conditional sentence. (4) […] that pun is in the Spanish and i think if you were only reading the English, you might just, you know think it's a typo. i don't know exactly what the translators should've done there, they could've called it fleambo instead of limbo or something i don't know, but anyway […] (LES300SU103) (5) […] but i don't know if that's the latest and greatest technology in uh i just know some systems used to work that way and i don't know if that's still the case, but horiz- the Chebyshev travel occurs whenever you have two independent […] (LES330JG052) Semantically, these verbs index judgment, ranging from speculating about the truth of the complement clause in one way or another, especially through mental and cognitive THE SEMANTIC-PRAGMATIC INTERFACE IBÉRICA 12 [2006]: 127-144 133 08 Artiga.qxp 20/09/2006 13:51 PÆgina 133 verbs3, to indicating deduction about the claim (guess, suggest, estimate, calculate …). Conveying certainty towards the truth of the proposition is illustrated in example 6 through an emphatic cleft sentence, whereas example 7 evinces how guess encodes speaker’s deduction on the upcoming information. It is also noticeable how these ELVs anticipate other lexical evaluators, particularly adjectives and adverbs (‘clearly’, ‘really’, ‘quite’, ‘important’, ‘good’ …) in both examples. (6) […] what i think that the the paper clearly does, is it shows that the kinds of sophistication that are in the option value or the stochastic dynamic program models, are really quite important […] (LES280JG138) (7) […] they're all equated with the Japanese syllabary, the Japanese Kana. uh there're forty-seven Kana. i guess that's a good way to remember how many Kana there are. (LEL140SU074) Apart from the recurrent mental process verbs, others describe more emotional or attitudinal states, particularly feel. It is interesting to note that feel conveys fuzzy connotations blending opinion and feeling as shown in example 8 where both meanings overlap. Besides, there is a further group of ELVs that, rather than marking the nature of the following sentence, provide evidence or report on the source of the evidence for the speaker’s comment, mainly appear, seem and show. This can be seen in example 9 where seem encodes the justification for the lecturer’s reasoning. (8) […] you don't have to say i love you all you have to do is give somebody food, and that indicates a relationship. (and) i feel that when i read this. (LES300SU103) (9) […] so, it seems to me that, if we understand the plot in relation to what the Spanish really says, it really means, when it happens you are bound forever. (LEL300SU076) The subject of these epistemic verbs controlling the complement clause can sporadically be explicit (10 times out of 375 in the case of think), although pronoun reference is much more frequent, being usually the first person singular, I, therefore the lecturer, who openly evaluates, judges or gives opinion. In the six most recurrent ELVs, I scores 330 out of 600 MARÍA ROSARIO ARTIGA LEÓN IBÉRICA 12 [2006]: 127-144134 Epistemic lexical verb I pronoun We pronoun You pronoun It pronoun think 256 13 13 4 not know 60 5 3 - seem 3 - - 20 suppose 1 1 4 6 assume 4 15 3 2 believe 6 2 2 1 TOTAL 330 55% 36 6% 25 4.1% 33 5.5% Table 3. Pronoun reference in lectures MICASE. 08 Artiga.qxp 20/09/2006 13:51 PÆgina 134 frequencies (55% of the total number of counts) as illustrated in table 3. Alternatively, yet not so frequently, we find the first person plural, we, by which speakers make a general reference also involving the audience (36 cases out of 600, 6%, in the six most frequent ELVs). Other personal pronouns include you, a clear indicator of the interpersonal function of language (25 counts, 4.1%, out of 600). You can show ambiguous reference but it usually relates to the audience and it is a further way for lecturers to bring students into the speech. In example 10 we can see instances of you (‘you may think’), explicit reference (‘your teacher may not think’) and also first person singular pronoun (‘I think’). Here think together with modals and adverbial modifiers (‘may’, ‘like’, ‘really’) indicates speculation. Impersonal pronoun reference (33 counts, a 5.5% of the total number of tokens) commonly correlates with seem (20 hits out of a total of 33 in the most common epistemic lexical verbs) and also appear (5 out of 10 hits). These copular verbs are usually followed by a subject predicative and perform a linking function so that the subject predicative directly characterizes the subject as seen in example 11. Seem can also enact the personal reference and subjective position by adding the object –to me as seen in example 12. Some of these verbs such as suggest, seem, appear, show, indicate and predict happily combine with both personal and impersonal pronoun reference. (10)[…] if you write a paper which is like graded, you know based on your, teacher's interpretation of your paper and you may think it's like really, mhm great and your teacher (may,) not think the same way and i think that that also causes like a little problem, cuz, as far as like, grades are concerned […] (LEL185SU066) (11)[…] and these matrix metalloproteases again help to break down the extracellular, matrix. it appears that both of these families of proteases they do somewhat different things and both of them are needed in order to get efficient […] (LEL175SU106) (12)[…] so, it seems to me that, if we understand the plot in relation to what the Spanish really says, it really means, when it happens you are bound forever. (LEL300SU076) Pattern 2: I + epistemic lexical verb (parenthetical sentences) Swales and Malczewski (2001: 149) refer to this pattern as ‘epistemic parentheticals.’ Parenthetical clauses are independent sentences which occur freely in different positions and interrupt the ongoing discourse at any point to qualify any part of the speech. Conrad and Biber (2000) even label them as ‘comment clauses’ acting as stance adverbials, used to “explicitly mark a proposition as the speaker’s opinion or to convey some level of doubt” (Conrad & Biber, 2000: 70). These authors contend THE SEMANTIC-PRAGMATIC INTERFACE IBÉRICA 12 [2006]: 127-144 135 08 Artiga.qxp 20/09/2006 13:51 PÆgina 135 that their use is limited to conversational registers and usually correlates with mental and cognitive verbs, especially think, believe, not know, feel, guess, imagine and suppose. Table 2 reveals 79 instances of parenthetical clauses (13.16% of the total number of tokens) in the six most frequent ELVs selected for the analysis. In monologic speech, as is the case in lectures, these units help break the flow of speech and allow speakers to express more tentative meanings. Example 13 shows speaker’s reservations on the information and in example 14 the lecturer hedges his/her claim in order not to appear too categorical about what he previously admitted; in both cases parentheticals enable non-commitment to propositions. (13)[…] now let me describe for you one, simple experiment, which makes this point. this is on the next page i believe of the coursepack. page, twenty-two. (LEL175SU106) (14)[…] trying to find out whether there's a relationship between birth control pill usage, and cancer risk, and i think i said there is no single experiment it's not like there's one right answer it's a very complicated question how scientists go about […] (LEL175SU106) Similarly, they can also be found at moments when lecturers hesitate or feel insecure in relation to the information transmitted, especially conveying complex conceptualizations in order to allow teachers time to find the best way to solve difficulties. In this sense, the use of ‘parentheticals’ can be associated with several hesitation pointers such as repetitions or other metatalk devices making linguistic chunks when the discourse is not smooth. In example 15 the speaker resorts repeatedly to this pattern as a hesitation marker at a moment of particularly difficult argumentation. (15)[…] overgeneral i thought. i mean, you can find other references, especially in the New Testament, i think that, that uh, indicate that it's a gift and i um, i think, no i don't think it i don't think it's i don't think it's looked down upon but i think in terms of um, i just i really don't like, um, when they say, it ju- i mean it's ar […] (LES605SU080) Pattern 3: impersonal/personal subject + epistemic lexical verb active/passive + clausal to infinitive Epistemic verbs are also frequently followed by a non-finite clause (62 out of 600 cases, a 10.33%, in the six most common ELVs, see table 2). However, this is restricted to seem and suppose in the selected corpus. This pattern rather than qualifying the proposition, evaluates the preceding subject, as can be seen in the examples below. The passive construction is slightly more common than its active counterpart (41 MARÍA ROSARIO ARTIGA LEÓN IBÉRICA 12 [2006]: 127-144136 08 Artiga.qxp 20/09/2006 13:51 PÆgina 136 tokens out of 600, a 6.83%, see table 2), being more formal and impersonal and making the attribution of the stance not explicit or ambiguous. Passive epistemic verbs enable speakers to either minimize full responsibility for their claims, protect their face or detach themselves from what is being talked about. The agent of the clause can be usually retrieved but, at times, it is difficult to decide on the hidden person behind the impersonal expression. Accordingly, this pattern tends to be preferred in those cases when lecturers present information that has not yet been proved or whose evidence is still unknown as in example 16 or when they refer to a general belief instead of manifesting his/her own views (example 17). In this example 17, the use of agentless passive patterns implies a lack of any specified human agency and contributes to impersonalization and a higher avoidance of any personal reference. (16)[…] this seems to be in contradiction, to another prescription. and the prescription that often run runs counter to to this, is foreign policy […] (LES495JU063) (17)[…] i didn't find very much um on the Christian views, of the Song of Songs but um it's supposed to be the romance of Jesus and the church with Mary as the chief female figure. (LES605SU080) Interestingly, both passive and active structures can also be sporadically accompanied by a passive infinitive adding further impersonalization to the utterance; something quite unusual in other conversational settings. Suppose is the verb that most commonly resorts to this typology in our corpus (‘is supposed to be supported’, ‘they are not supposed to be finished’, ‘was supposed to be chosen’ …). In example 18, the unknown and blurred attribution is indexed to speculate about the content without clearly indicating the source of the judgment. (18)[…] i mean the countries that were most affected by those missiles that were supposed to be deployed in Western Europe, um, England, there were four countries England Netherlands Germany and Italy. (LES495JU063) Pattern 4: noun/adjective/verb + epistemic lexical verb infinitive + that-clause Another recurrent pattern containing ELVs is that of the infinitive form after either a noun (‘temptation’, ‘tendency’, ‘reason’ …), an adjective (‘likely’, ‘interesting’, ‘difficult’ …) or even other alternatives such as a verb followed by infinitive (‘have’, ‘make’, ‘want’ …). This pattern accounts for 37 cases (6.16%) in the six most common ELVs in our data as seen in table 2. In the example below we can observe how the epistemic lexical infinitive verb qualifies the upcoming that-clause proposition by making a hedged general assumption the lecturer disagrees with. THE SEMANTIC-PRAGMATIC INTERFACE IBÉRICA 12 [2006]: 127-144 137 08 Artiga.qxp 20/09/2006 13:51 PÆgina 137 (19)[…] so, there's always this temptation to think that evolution is moving us, in a direction towards always higher performance that's not what evolution is doing at all. evolution is a pr […] (LEL500JU034) As regards the phraseology these stance markers usually reflect, it is recurrent to find different types of hedging, including prepositions such as ‘like’ and adverbs ‘sort of ’ and ‘just’, either following or preceding the epistemic verbs making sentences blurred or more fuzzy and adding ambiguity to the speaker’s position. These associations, typically restricted to informal conversational settings, are frequent with feel and seem as illustrated in example 20, where reformulation and the epistemic modal and adverbs contribute to further tentativeness. (20)[…] and even though some of what they produce might seem like, it's you know geared to, uh, disaffected women, or people of color or seeks to show, uh poor people a little bit. that that's all kind of a mask. (LEL220JUO71) Evidence makes it also clear that these lexical verbs do not often occur in isolation, on the contrary, they seem to cluster with other stance devices including prosodic features, discourse fillers, evaluative adjectives, adverbs and modality markers, especially modal verbs, shaping mutually supporting associations that allow speakers to index more epistemic meanings, usually making sentences less likely and more tentative. In example 21 we can see suggest preceding two evaluative adjectives. The most commonly found modal verb in our corpus is the conversational ‘can’, which usually collocates with imagine and think, another typical association is ‘should’ and know. ‘May’, ‘might’ and ‘would’ are currently found with think, seem, suggest, suppose … hedging the discourse, as seen in example 22 above. (21)[…] when she goes into the dark room that is no longer dark, with Pedro. in fact, i would suggest to you that, one of the great things about this novel is that it works with an enormous set of oppositions. (LES300SU103) (22)[…] if you write a paper which is like graded, you know based on your, teacher's interpretation of your paper and you may think it's like really, mhm great and your teacher (may,) not think the same way and i think that that also causes like a little problem […] (LEL185SU066) Overall, it is frequent to find high traffic of these epistemic lexical verbs making bundles as observed in examples 23 and 24; this proliferation of stance markers allows speakers to achieve their purposes more effectively, not only to facilitate lecture comprehension but also to soften power status inequalities between speakers and audience. This feature correlates with mechanisms of spoken registers and includes retrace and repair, reformulation, clarification, pauses and other hesitation markers, characteristic of improvised speech. MARÍA ROSARIO ARTIGA LEÓN IBÉRICA 12 [2006]: 127-144138 08 Artiga.qxp 20/09/2006 13:51 PÆgina 138 (23)[…] so along with stock options which i think is a great idea, i think, you know, some kind of like, the company should_ i mean for_ i guess Brenda is not in control other than uh, you know, re-saying to the company what she thinks is uh_ somehow getting a company, organization […] (LEL185SU066) (24)[…] especially in the New Testament, i think that, that uh, indicate that it's a gift and i um, i think, no i don't think it i don't think it's i don't think it's looked down upon but i think in terms of um, i just i really don't like, um, when they say, it ju- i mean it's archaic language obviously but when they say he knew he lay with her it just seems like, a chore […] (LES605SU080) In the next section, attention will be focused on the contextual use of think, the most representative epistemic lexical verb in the corpus (51.08%, more than half of the total number of occurrences and a frequency of 24.5 items per 10,000 words) to interpret the functional variability and pragmatic implications of this epistemic marker within the context of academic lectures. Pragmatic dimension of epistemic lexical verbs Think emerges as the most salient epistemic lexical verb marker in our corpus and performs several communicative functions. Primarily, think is a recurrent epistemic stance marker and commonly indexes the speakers’ opinion about the content, a characteristic function of academic speech where lecturers, as members of a discourse community, indicate degree of certainty, comment on the evidence, assess knowledge, or commit themselves towards the information presented in a more subjective and persuasive manner. This position of stance is also associated with personal authority and power but since speakers limit their claims to their personal opinion they also give their audience freedom to refute or disagree with them. Therefore, the principle of solidarity usually reflected in this instructional genre (Swales, 1990) is not disrupted. In examples 25 and 26 speakers give their own tentative views about the content. On some occasions, meanings are blurred and it is not clear whether the utterance is a declaration of opinion or of uncertainty, as in example 26. (25)[…] the meaning of the metaphor with um the rabbinic metaphor or a feminist metaphor or if there's like, a mix of both. i think i think it probably could be a mix of both like, i think both interpretations are valuable i think like, the feminist interpretation is […] (LES605SU080) (26)[…] make the bass rise by half steps until he gets there. okay same kind of thing going on here only i don't think it's very funny. i don't think it would be much exaggeration at all to say that that one interval the rising half step is the basis of the development section. (LES420MG134) THE SEMANTIC-PRAGMATIC INTERFACE IBÉRICA 12 [2006]: 127-144 139 08 Artiga.qxp 20/09/2006 13:51 PÆgina 139 At other times, rather than expressing evaluation towards content, think acts as an attitude marker to convey more personal views and feelings. This function is typical of conversational exchanges but it is also present in academic settings. Through attitude markers lecturers rather than conveying judgment towards content, show alliances with the information. Speakers make use of interpersonal resources to establish relations with audiences and index their own point of views, which promotes conviviality. We can also note that this verb in the affirmative encodes some kind of agreement or certainty whereas disagreement and uncertainty correlate more with negative forms. According to Nattinger and DeCarrico (1992), ‘I think’ is one of the most frequent lexical phrases used in lectures for signaling agreement and disagreement. In example 27 uncertainty seems to be close to criticism towards previous views on the Bible whereas example 28 reveals more some degree of likelihood. In both examples think anticipates lexical markers of evaluation. (27)[…] ch on the rabbinic um, research of it. but um ... y- no i i don't think that the Bible is wholly again- anti, sexuality. but i do think that, um, it's, very cautious maybe even overly cautious. in a modern sense. (LES605SU080) (28)[…] okay. uh he's a lotta fun. i think i think Ueda Akinari's interesting. uh cuz his plots are so interesting they have such interesting twists. uh House Amid the Thickets […] (LEL140SU074) A further functional variable of “think” corresponds to that of metadiscourse device to organize and signal discourse conveniently as suggested by Schiffrin, 1980; Crismore et al., 1993; Mauranen, 2001; Swales and Malczewski, 2001, and Simpson, 2004, among others. These discourse management elements are part of the rhetoric of this genre and proliferate in pedagogic settings where an effective transmission of information is aimed at. Again, speakers resort to the interpersonal elements of language to establish rapport with audiences about the upcoming speech, draw attention towards some specific aspects or demand participation and feedback. In example 29 think appears as an attention-getter to make things salient to students and in example 30 it is used to seek for students’ participation by eliciting their response. In both instances think forms part of metadiscourse elements enhancing teacher classroom control. (29)[…] Tita Chencha and John dried the bedroom, the stairs, and the bottom floor. and so it goes from stairs to stairs. i think that this section is really a paradigm for the way the whole book functions. (LEL300SU076) (30)[…] hold it within the vesicle so that it can't be targeted ...? you're whispering the answer, what? whatta you think? no, you don't know? the cholesterol for (xx) yes, L-D-L receptors. (LEL175SU106) A recurrent pragmatic functional category of think in the corpus is that of hedging the discourse in various ways. Hedges (Hyland, 1996, 1998; Markkanen & Schröder, MARÍA ROSARIO ARTIGA LEÓN IBÉRICA 12 [2006]: 127-144140 08 Artiga.qxp 20/09/2006 13:51 PÆgina 140 1997) constitute rhetorical elements aimed at avoiding categorical claims and diminishing the force of bald assertions which fit perfectly in spoken genres. It is usually hard to pinpoint the exact purpose of hedges since their meanings commonly overlap but they are frequently intended to: � show cautious positions and tentativeness, allowing students to contradict speaker’s opinions (e.g. ‘I think this is a kind of complicated uh uh statement’; ‘I think it’s not perhaps quite as clear as in the English’; ‘I don’t think it’s looked down upon’ …) � make criticism smoother (e.g. ‘your statement was a little, I thought, overgeneral’; ‘I think you’re grown ups, some of you, I think, have gone to great pains and at great length’ …) � give advice or recommendations that would otherwise sound too pushy (e.g. ‘I don’t think I have to read it verbatim’; ‘I think you would much more quickly get more’ …) � show sympathy and agreement (e.g. ‘I think what you are saying makes sense’ …) � simply appear vague or ambiguous (e.g. ‘I think that’s sort of what is going on’; ‘we think we know what’s real what reality is, something which has a little bit of I would think a folksy element’ …) In truth, we observe in the corpus that lecturers generally adopt positions of modesty; therefore, most of the meanings of hedges can be related to politeness (Brown & Levinson, 1987) enabling speakers not to impinge on the negative face of audiences by manifesting too harsh assertions. In this example the lecturer’s avoidance of categorical positions is evinced by the use of several hedges additionally downtoned by modals and adverbs. (31)[…] that pun is in the Spanish and i think if you were only reading the English, you might just, you know think it's a typo. i don't know exactly what the translators should've done there, they could've called it fleambo instead of limbo or something […] (LES300SU103) Finally, think appears as a discourse filler or floor maintaining filler (Poos & Simpson, 2002) at moments of hesitation, allowing speakers time to regain the flow of speech or at times when the transmission of information is not smooth. In this function think usually clusters in bundles with other epistemic markers and with additional discourse markers. Indicators of spontaneous speech including pauses, reformulation, retrace and repair, stammering and repetitions are frequent as we can see in example 32 where the speaker seems to have probably lost the train of thought. THE SEMANTIC-PRAGMATIC INTERFACE IBÉRICA 12 [2006]: 127-144 141 08 Artiga.qxp 20/09/2006 13:51 PÆgina 141 (32)[…] my question actually. um, i think i think that she's talking, i mean it's i think she's talking um, i think it's the bride and her friends are talking, uh, or um, her ... i think it's her friends actually, talking to her. right. (LES605SU080) Undoubtedly, since stance markers overlap and bundle, it is particularly difficult in most cases to establish speakers’ real intentions and to pinpoint on the specific function think performs in the discourse. It is therefore more accurate to conclude that it is a multifunctional element to convey several communicative meanings. Concluding remarks By tracing recurrent phraseological units in academic speech, corpus evidence showed that university lectures do not seem to be neutral or objective. On the contrary, this genre appears to be persuasive and highly evaluative making frequent use of ELVs, among other devices, to index speakers’ personal stance. Data suggested that among the items under study, judgmental verbs especially think were the most recurrent followed by much more sporadic elements such as not know, seem, suppose, assume and believe. Also, in spite of the freedom and openness that characterizes the syntax and lexis of conversation, we found patterns associated to meanings. These verbs form distinctive lexical structures of use and recurrent patternings that correspond to semantic meanings. Particularly salient was the use of these epistemic lexical verbs to frame that-clauses whose propositions they evaluate and to introduce other lexical markers of evaluation, mostly adjectives and adverbs. There was also a correlation of this structure with first person pronouns that make opinions more prone to criticism and give the discourse a flavor of subjectiveness. The analysis also showed how these units cluster at moments making bundles to attain their purposes more effectively. From a pragmatic viewpoint, ELVs have shown to function as rhetorical devices that help connecting language to producers establishing participatory relationships with audiences in various different ways. The pragmatic effects of stance as a linguistic enactment of interpersonal meanings have included in the analysis above evaluating messages, marking difficulties in the transmission of information, hedging the discourse and complying with politeness conventions. Lecturers rather than adopting categorical positions seemed to manifest their views and project their own persona in a mitigated way, encouraging solidarity and fostering a participatory and convivial atmosphere within the classroom. On pedagogical grounds, this small scale study has shed some light on the convenience of instructing students into recurrent phraseology and into the MARÍA ROSARIO ARTIGA LEÓN IBÉRICA 12 [2006]: 127-144142 08 Artiga.qxp 20/09/2006 13:51 PÆgina 142 THE SEMANTIC-PRAGMATIC INTERFACE IBÉRICA 12 [2006]: 127-144 143 pragmatic meanings ELVs convey. Moreover, the study might make us think of this category of lexical verbs as an essential open typology to mark stance positions that should be paid attention to in the curricula. In all, non-native speakers should be aware of the role stance markers perform in order to avoid pragmatic failure and being at the risk of sounding too categorical or impolite in their exchanges. Acknowledgments Thanks are due to Dr. G. R. Ferguson (University of Sheffield) for his sound advice on the analysis of ELVs and to Dr. C. Pérez-Llantada (University of Zaragoza) for her comments on previous drafts of this paper. I also want to express my gratitude to the anonymous reviewers of Ibérica for their helpful suggestions. (Revised paper received May 2006) Barton, E. (1993). “Evidentials, argumentation, and epistemological stance.” College English 55: 745-69. Beach, R. & C.M. Anson (1992). “Stance and intertextuality in written discourse.” Linguistics and Education 4: 335-57. Biber, D., & E. Finegan (1988). “Adverbial stance types in English.” Discourse Processes 11: 1-34. Biber, D., & E. Finegan (1989). “Styles of stance in English: lexical and grammatical marking of evidentiality and affect.” Text 9: 93-124. Biber, D., S. Johansson, G. Leech, S. Conrad & E. Finegan (1999). Longman Grammar of Spoken and Written English. Essex: Longman. Brown, P. & S. Levinson (1987). Politeness: Some Universals in Language Usage. Cambridge: Cambridge University Press. Chafe, W. L. & J. Nichols (eds.) (1986). Evidentiality: The Linguistic Coding of Epistemology. Norwood, NJ: Ablex. Conrad, S. & D. Biber (2000). “Adverbial marking of stance in speech and writing” in Hunston & Thompson (eds.), 56-73. Crismore, A., R. Markkanen & M. Steffensen (1993). “Metadiscourse in persuasive writing: A study of texts written by American and Finnish university students.” Written Communication 10,1: 39-71. Givón, T. (1993). English Grammar: A Functional-based Introduction, vol. 2. Amsterdam: John Benjamins. Halliday, M. A. K. (1994). An Introduction to Functional Grammar. London: Arnold. Holmes, J. (1988). “Doubt and certainty in ESL textbooks.” Applied Linguistics 91: 20-44. Hunston, S. (1994). “Evaluation and organization in a sample of written academic Discourse” in M. Coulthard (ed.), Advances in Written Text Analysis, 191-218. London: Routledge. Hunston, S. & J. Sinclair (2000). “A Local Grammar of Evaluation” in Hunston & Thompson (eds.), 74-101. Hunston, S. & G. Thompson (eds.) (2000). Evaluation in Text. Authorial Stance and the Construction of Discourse. Oxford: Oxford Linguistics. Hyland, K. (1996). “Writing without conviction? Hedging in science research articles.” Applied Linguistics 17,4: 433- 454. Hyland, K. (1998). Hedging in Scientific Research Articles. Amsterdam: John Benjamins. Markkanen, R. & H. Schröder (1997). Hedging and Discourse. Approaches to the Analysis of a Pragmatic Phenomenon in Academic Texts. Berlin: Walter de Gruyter. Martin, J. R. (2000). “Beyond Exchange: Appraisal Systems in English” in Hunston & Thompson (eds.), 142-175. Mauranen, A. (2000). “‘A Good Question.’ Expressing Evaluation in Academic Speech.” Paper presented at the 5th ESSE Conference, Helsinki, August 2000. Mauranen, A. (2001). “Reflexive Academic Talk: Observations from MICASE” in R.C. Simpson and J.M. Swales (eds.), Corpus Linguistics in North America. Selections from the 1999 Symposium, 165-178. Ann Arbor: University of Michigan. Nattinger, J. R. & J. S. DeCarrico (1992). Lexical Phrases and Language Teaching. Oxford: Oxford University Press. REFERENCES 08 Artiga.qxp 20/09/2006 13:51 PÆgina 143 María Rosario Artiga León is a teacher of English at the Escuela Oficial de Idiomas “Lázaro Carreter” of Zaragoza and a PhD student at the English Department of the University of Zaragoza. She is currently working on a corpus-based thesis about stance in academic speech. Her main research interests are pragmatics, ESP, EAP, corpus linguistics and applied linguistics. MARÍA ROSARIO ARTIGA LEÓN IBÉRICA 12 [2006]: 127-144144 Nöel, D. (2002). Believe-type Matrix Verbs and their Complements. Corpus-based Investigations of their Functions in Discourse. Gent: University of Gent. Ochs, E. (ed.) (1989). “The Pragmatics of affect.” Special issue of Text 9,3: 7-25. Pawley, A. & F. Syder (1983). “Two puzzles for linguistic theory: nativelike selection and nativelike fluency” in J. C. Richards & R. W. Schmidt (eds.), Language and Communication, 191- 226. New York: Longman. Poos, D. & R.C. Simpson. (2002). “Cross-Disciplinary Comparisons of Hedging: Some Findings from the Michigan Corpus of Academic Spoken English” in R. Reppen, S. M. Fitzmaurice & D. Biber (eds.), Using Corpora to Explore Linguistic Variation, 3-21. Amsterdam/ Philadelphia: John Benjamins. Schiffrin, D. (1980). “Metatalk: Organisational and Evaluative Brackets in Discourse.” Sociological Inquiry 50: 199-236. Scott, M. (1999). Wordsmith Tools, Version 3.0. Oxford: Oxford University Press. Simpson, P. (1993). Language, Ideology and Point of View. London: Routledge. Simpson, R. C. (2004). “Stylistic Features of Academic Speech: The Role of Formulaic Expressions” in U. Connor & T. A. Upton (eds.), Discourse in the Professions: Perspectives from Corpus Linguistics, 37-64. Amsterdam: John Benjamins. Sinclair, J. M. (1991). Corpus Concordance Collocation. Oxford: Oxford University Press. Swales, J. M. (1990). Genre Analysis: English in Academic and Research Settings. Cambridge: Cambridge University Press. Swales, J. M. & B. Malczewski (2001). “Discourse Management and New-Episode Flags in MICASE” in R. C. Simpson & J. M. Swales (eds.), Corpus Linguistics in North America. Selections from the 1999 Symposium, 145-164. Ann Arbor: University of Michigan. Swales, J. M. & A. Burke (2003). “‘It’s Really Fascinating Work’: Differences in Evaluative Adjectives across Academic Registers” in P. Leistyna & C.F. Meyer (eds.), Corpus Analysis. Language Structure and Language Use, 1-18. Amsterdam – New York: John Benjamins. 1 Hyland (1998) and Biber et al. (1999) agree that think, know, seem, suppose, assume and believe are recurrent in conversational settings. 2 According to MICASE index of transcripts, the lectures used for this research correspond to the following file identification and file names: LEL115JU090: Intro Anthropology Lecture LEL140SU074: Japanese Literature Lecture LEL175SU106: Biology of Cancer Lecture LEL185SU066: Behavior Theory Management Lecture LEL220JUO71: Intro Communication Lecture LEL300SU076: Fantasy in Literature Lecture LEL500JU034: Intro Psychology Lecture LES280JG138: Labor Economics Lecture LES300SU103: American Literature Lecture LES330JG052: Graduate Industrial Operations Engineering Lecture LES420MG134: Beethoven Lecture LES485MG006: Graduate Physics Lecture LES495JU063: Political Science Lecture LES605SU080: Women in the Bible Lecture 3 These mental and cognitive verbs are the most frequent in our corpus and include think, believe, assume, suppose, suspect... NOTAS 08 Artiga.qxp 20/09/2006 13:51 PÆgina 144