Epistemicity and stance: A cross-linguistic study of epistemic stance strategies in journalistic discourse in English and Spanish

Epistemic stance

Epistemic stance acts pertain to the positioning of the speaker/writer with respect to knowledge about described events and their commitment to the validity of the communicated information. They are characteristically aimed at the legitimization of the assertions, through the expression of speaker/writer’s degree of certainty regarding the realization of the event and/or the reference to the sources and modes of access to that knowledge. The conceptual domain of epistemicity, as Boye (2012) consistently argues, comprises the ‘subcategories evidentiality and epistemic modality’ (pp. 2–3). Evidentiality is defined as ‘epistemic justification’, in terms of notions such as ‘source of information’, ‘evidence’ or ‘justification’ (pp. 2–3). Epistemic modality is conceived as ‘epistemic support’ and involves notions such as ‘degree of certainty’ and ‘degree of commitment’ (pp. 2–3).

Evidentials primarily indicate the source of information and the evidence on the basis of which the speaker feels entitled to make a claim (Anderson, 1986). As Anderson (1986) notes, ‘Evidentials denote the kind of “justification” a speaker has for making a claim, whether that comes from direct evidence or observation, evidence plus inference, inference with unspecified evidence, an expectation from logic, or whether the evidence is visual, auditory, etc’ (p. 274). A broader conception of evidentiality includes both the source of information and an estimation of its reliability, as Chafe (1986) posited in his seminal publication.

Within the literature on evidentiality, the focus has traditionally been on grammatical evidentiality, that is, those systems of languages where marking the information source is obligatory and where evidentiality is a grammatical category in its own right (Aikhenvald, 2004). Aikhenvald (2007) also notes that ‘Every language – no matter whether it has a fully grammaticalized evidential system, or evidential extensions of other categories – has an array of further ways of describing how one knows things’ (p. 209). She argues for a distinction between the term ‘evidentiality’ to refer to the grammatical category of marking information source and the corresponding conceptual category ‘information source’: ‘the term “evidential” is best used for closed grammatical systems, and the term “information source” for the vast body of other ways of referring to knowing things’ (Aikhenvald, 2007: 222). In Germanic and Romance languages, like English or Spanish, which do not have obligatory grammatical evidentiality, we find a broad array of expressions, open lexical classes (verbs, adverbs and adjectives), closed subclass of modal verbs, parentheticals, particles and various evidential constructions, which share a substantial range of semantic extensions to signal the information source, and that shared phenomenon begs explanation from a functional-onomasiological perspective (cf. Wiemer, 2010). Thus, if we restrict the notion of evidentiality to cases of obligatory grammatical marking, we miss out on the expression of evidentiality in a significant number of languages, and we fail to adequately characterize and delimit the conceptual domain of evidentiality, or ‘information source’ to use Aikhenvald’s (2007) term, in these languages. As Lampert and Lampert (2010) argue, ‘the category of evidentiality is of use only, we conjecture, if a radical conceptual stance is taken in order to not miss capturing alternative linguistic strategies of expressing this notion. Hence, we strongly endorse the primacy of functional criteria and suggest including all linguistic representations that serve as cues for evidentiality in context’ (p. 319). Perhaps the term ‘evidential expressions’ might contribute to establish a distinction between this broader category of expressions signalling the conceptual domain of ‘evidentiality’, which at the same time also convey a wider and more fine-grained range of meanings (belief state or attitude to what is said) related to ‘subjectification’ (Traugott, 1995), as Aikhenvald (2007: 220) argues, in contrast with the grammatical category of evidentiality which merely signals type of source.

In the literature, we find various subdivisions of the domain of evidentiality, such as Diewald and Smirnova’s (2010: 74) classification of evidential values and similar distinctions adopted within the EUROEVIDMOD project (Wiemer and Stathi, 2010). In this article, we restrict our focus to two of the values or subcategories of evidential expressions: (a) indirect-inferential evidence (IIE) and (b) indirect-reportative evidence (IRE).

Expressions of IIE include those markers which indicate both perceptual-based and conceptual-based inferences, as well as inferential meaning based on information acquired through oral and written reports (report-based inferences), as the following examples show:

(1) IIE (+Perceptual): Members of the Taleban and al-Qaeda were filmed kneeling in the dirt, their eyes shielded by goggles and thick black tape, their ears covered with muffs and their mouths with surgical masks. Chained to each other and restrained by leg irons and handcuffs, the prisoners look like helpless victims. The photographer appears <IIE> to have caught their captors in the course of a gross violation of human rights. (ELT ¹ )

(2) IIE (+Perceptual): El resultado de ello es tan elocuente como abrumador: nadie puede permanecer indiferente a una lectura serena y atenta del texto y de los centenares de ilustraciones que lo acompañan: las hienas, monos y espantajos reproducidos en todas las formas y tamaños, ¿pertenecen a la especie humana? Aparentemente <IIE>, no. (SOP)

     (… Do these (forms and shapes) belong to the human species? Apparently, not.)

(3) IIE (+Conceptual): A burst of publicity can have as heady an effect on politicians as a bottle of champagne. It can also leave them feeling rotten the next day. William Hague went repeatedly for the high and let himself in for the subsequent hangovers, but I had thought Iain Duncan Smith was better able to resist. It seems <IIE> I was wrong. The Conservative leader’s hijack of the case of Rose Addis has won him a few days’ headlines. His colleagues are pleased that he has made an impact. But at what cost? (EOT)

(4) IIE (+Conceptual): Que haya fricciones entre vecinos es la consecuencia lógica cuando uno de ellos no hace todo lo que podría para llegar a una situación de armonía. Las reclamaciones de España no sólo no han sido atendidas como debieran, sino que da la impresión <IIE> de que Rabat hace la vista gorda como método de presión, especialmente en lo relativo a la oleada migratoria y al cultivo y tráfico de hachís. (SLA, ABC)

 (The complaints posited by Spain have not only not being attended to as they should, but rather it gives the impression that Rabat is turning a blind eye as a form of pressure, …)

A distinction is drawn here between cases which involve report-based inferences and would thus be classified together with other forms of IIE, as in the following examples, and those cases of IRE where no internal inferential process is involved:

(5) IIE (+Report): Damp squib or smoking gun? The conclusions from Lord Butler’s report seem <IIE>, at first sight, to provide the expected establishment support for the government. Lord Butler proves himself no patsy Hutton. He’s shrewder than the judge, a real player. He makes several trenchant criticisms. But he decides, in the end, that no one’s actually to blame. (EOG)

(6) IIE (+Report): No he leído todavía la letra pequeña de la sentencia, pero parece <IIE> que queda claro que ni ETA tuvo nada que ver con tan lamentable masacre ni se observa en su fondo el más mínimo rastro que permita establecer las relaciones causa-efecto que han alimentado el amarillismo periodístico. (SOA)

 (I have not yet read the details of the sentence, but it seems that it is clear that neither did ETA have anything to do with that awful massacre …)

Expressions of IRE involve non-personal, indirect, external access to the evidence through other speakers/writers’ reports. Very often there is no reference to the actual source or external voice of the reported information:

(7) IRE: In Italy, a judge is considering whether there is enough evidence to try 26 Americans and six Italians for their role in the 2003 alleged kidnapping of Muslim cleric Hassan Mustafa Osama Nasr, known as Abu Omar. As a result of this reaction across Europe against rendition, CIA agents are reportedly <IRE> increasingly taking out their own private liability insurance to protect themselves from prosecution abroad. (EOT)

In Spanish, especially in the context of journalistic discourse, there is an increasing tendency to use the conditional perfect form of the verb (habría/n) with an evidential reportative value in news reports on the radio or on TV (personal observation). This use has relatively recently permeated into the written discourse and appears to be increasing during the past few years (personal observation). In our corpus of journalistic discourse ranging from 1999 to 2010, we have found 16 tokens, mostly in news reports, as in the following example:

(8) IRE: Hobeika habría <IRE> dejado grabaciones y pruebas que implican a Sharon en la masacre de Sabra y Chatila (SNA)

 (Hobeika had reportedly left some recordings and some proof which implicated Sharon in the massacre in Sabra and Chatila)

Epistemic modals (EMs) pertain to the conceptualizer’s striving for control of conceptions of reality (Langacker, 2009, 2013). They express the speaker’s assessment concerning the existence of the event designated and its actual or potential occurrence. As Langacker (2008) observes, ‘the absence of a modal indicates that C accepts the profiled process as being real’, as part of the conceptualizer’s conceived reality (Rc), whereas the presence of a modal places the profiled process ‘outside conceived reality, in a region we can refer to as irreality (the complement of Rc)’ (p. 302). EMs reflect different degrees of epistemic support regarding the realization of the event designated in a proposition, that is, full support (total certainty), partial support (probability) and neutral support (possibility and uncertainty) (Boye, 2012):

(9) EM: The Bush camp can argue that pessimists have been predicting apocalypse for years and it hasn’t happened. This is true. It may <EM> not happen for another few years either; but equally the fall from grace could <EM> be quite soon. Once markets change their minds, their revenge can be swift and apolitical, as John Major’s administration found to its cost. (ELG)

(10) EM: Puede <EM> que muchos indecisos se hayan inclinado finalmente por Bush asustados por todos quienes han formado el frente antirrechazo al mismo. Desde el patético Michael Moore al joven socialismo español, pasando por Chirac y medio Hollywood. (SOA)

       (Many undecided voters may have finally decided to back up Bush …)

Within epistemic stance resources, this article also includes verbs of mental state or cognitive attitude (CGA), which indicate speakers’ reflective attitudes or beliefs regarding representations. As Capelli (2007) observes, these expressions involve the epistemic evaluation or ‘the assignment of a “likelihood degree” to a representation’, that is, ‘when the evaluator is certain that a representation of some sort holds, the output is a piece of knowledge, otherwise when the degree is still positive but below the level of certainty, the output is a belief’ (pp. 123–124):

(11) CGA: Yet as he goes, it seems to me <CGA> that the biggest gap between Blair and many of us who have watched him for years, is the same gap as we first noticed at university. It’s the fact of his committed Christianity, and the mystery (to me) of how that affects his self-certainty. (EOG)

(12) CGA: Cuando muchos, durante la instrucción y la posterior fase de juicio oral en el macroproceso sobre el 11-M, afirmaban que la sentencia nos proporcionaría la verdad acerca de una de las matanzas terroristas más atroces cometidas en el mundo occidental, me parece <CGA> que estaban mostrando un gran desconocimiento (o quizá consciente desprecio) de lo procesal y de la experiencia histórica. (SOA)

      (…, it seems to me that they were showing a great lack of knowledge …)

In addition to the values basically attached to particular evidential expressions, we find that certain evidential expressions undergo semantic extensions in their signalling of information sources, which are discussed in the following section on the phenomenon of multifunctionality.

Multifunctionality of evidential expressions

On the basis of evidence from a sample of 50 languages of a different genetic affiliation, Boye (2012: 137ff) observes that there are a number of expressions that are synchronically polyfunctional or others that are moving diachronically, with respect to specific subspaces or notional regions within the semantic map of epistemicity, which would seem to point to certain bidirectional connecting links between those spaces. Within the domain of evidentiality, this link is manifested in the use of predicates of visual perception, such as see or ver, which may be used with the value of direct evidence and undergo extensions into IIE:

(13) IIE: The first is simply a return to the normal sovereignty of the nation states. That is the default setting for Europe, as one can [[see]] from the condition of the single currency. If the euro fails, most of the eurozone nations will withdraw; they will return to their original independence and their original currencies, though … (EOT)

(14) IIE: Ese mismo sustrato de insensibilidad hace que el PSOE incumpla descaradamente una promesa electoral sin que pase nada. Porque se comprometieron a elaborar una ley marco de protección animal, pero el Gobierno acaba de declarar que no la hará y que las competencias son de las autonomías (se han presentado 1.300.000 firmas en pro de la ley, pero [[se ve]] que les importa un pito). (SOP)

 (… Because they promised to pass a law on protection of animals, but the Government has just declared that it won’t do so and that the competency belongs to the autonomous governments (1,300,000 signatures in favour of the law have been presented, but it seems that they don’t give a damn).)

Within the domain of epistemic modality, a connecting line is found between full support or certainty meaning and partial support or probability, reflected, for example, in the diachronic movement of French sans doute, which has now come to express ‘less than full support’ (Boye, 2012: 141). This devaluation is also apparent in English ‘no doubt’, where we find examples which are closer to a probability meaning:

(15) CGA: Shortly before the 1997 election I had to inform my boss that I was pregnant, though no doubt <CGA> he’d guessed since I’d waited almost 20 weeks, hiding my girth under Ruth Badger-style jackets, so terrified was I of his response. (EOT)

Cross-domain connecting lines between semantic spaces within evidentiality and those within epistemic modality are also found. As Boye (2012) notes, multifunctional expressions covering direct justification and full support or epistemic certainty are attested, for example, in some Polynesian and Australian languages, where a single perception verb covers the meanings of ‘to see’ and ‘to know’ (Viberg, 1984). Finally, both synchronic polyfunctionality and diachronic changes support a link posited between indirect justification or evidence and partial support or epistemic probability.

Boye (2012: 157) notes that a developmental path from full support to indirect justification has not been independently attested. However, an interesting example of what might constitute an extension from full support to indirect justification is the case of the expression ‘se conoce’, from the verb ‘conocer’ (know), as can be observed in the following example, which might be deemed a case of strong partial support:

(16) CGA/IIE: Steiner sostiene que no se puede ser profesor sin ser por dentro un déspota, pensando: «Te voy a hacer amar un texto bello, una bella música, las altas matemáticas, la historia, la filosofía…» Y advierte de la extrema ambigüedad de la ética que sostiene a esta esperanza. Cómo [[se conoce]] que no conoce la fuerza retórica de esos españoles que, como Badiola, el veterinario, quisieron cambiar el país un día y lo lograron. (SOA)

 (… And [Steiner] warns against the extreme ambiguity of the ethics that sustains that hope. How obvious it is that he does not know the rhetorical force of those Spaniards, who, like Badiola, the veterinarian, wanted one day to change the country and made it.)

Diachronic extensions of IIE value have been posited for epistemic necessity modals, English must and Spanish deber (de) (Van der Auwera and Plungian, 1998). This extension would involve a link between partial support and indirect justification. However, this evidential extension, as Boye (2012) notes, is ‘not entirely uncontroversial’ since it would imply a prior epistemic modal extension where ‘non-epistemic necessity extends into partial, rather than full support’ (p. 153). As Boye (2012) observes, ‘it might turn out after a detailed analysis that only strong partial support, as opposed to weak partial support, is linked to indirect justification’ (p. 158):

(17) EM/IIE: For Mr Blair, who presides over a party still riven by much deeper ideological fissures than were Mr Major’s Tories, running a Government with a modest majority must <EM> be a daunting prospect. (EOT)

(18) EM/IIE: … después de este paréntesis, la representación de Cataluña irá a dar en quien de verdad la merece que es Carod Rovira, aunque ya no será él, sino alguno de sus amigos y colegas que debe de <EM> estar segándole la hierba bajo los pies. Es lo justo para Cataluña, aunque signifique la muerte de Barcelona. (SOP)

      (… but one of his friends who must be mowing the grass beneath his feet …)

There has traditionally been certain controversy in the literature regarding the status of English EM must. Van der Auwera and Plungian (1998) note that English must is best situated in the overlap between the semantic space of inferential evidentiality and that of epistemic necessity. According to the cognitive linguistics conception of epistemic modality in terms of force dynamics, this meaning extends from the notion of force (compulsion vs the absence of a barrier) attached to ‘root modals’ must and may manifested at the social level to the mental level (cf. Sweetser, 1982; Talmy, 1988). In the case of EMs, Langacker (1991) conceives this potency as inherent in the conceptualizer’s mental activity, in their assessment of the ‘evolutionary momentum of reality’: in its epistemic value, must situates the designated process in projected ‘known reality’ (p. 274). Langacker (1991) points out that ‘modal senses based on the notion of evolutionary momentum might well engender the conception of the speaker being carried by the force of evidence along a deductive path’ (p. 274). The notion of potency is thus interpreted at the mental level as the force of evidence which leads the speaker to the most likely conclusion. Similarly, Lampert (2011) notes that

In definitions of epistemic modality in terms of a mapping of forces, the notion of evidence becomes part of the definition, thus making obvious the relatedness of the two concepts, but at the same time also the problems involved; in the epistemic domain, the only imaginable ‘force’ can be the ‘force of evidence’ (in the form of premises), which would, for instance in the case of epistemic must, compel the speaker to conclude that p. But the force of evidence is also present in the other epistemic modals, such as may and might – the only difference between them perhaps being that the conclusion is stronger or weaker. (p. 9)

De Haan (2009) argues that the conclusional force of must does not necessarily imply that it should be categorized as an evidential marker and calls for a distinction between those elements that ‘simply assert evidence’ and those which ‘evaluate evidence (according to its likelihood of being true)’ (p. 268). In the words of De Haan (2009):

While it is true that must can be analyzed as a modal that draws conclusions from evidence […], this does not automatically mean that must is an evidential. It is not clear that evidentials draw conclusions from evidence. Rather, evidentials show that there is evidence for the statement the speaker is making, without drawing any (modal) conclusion from the evidence. (p. 268)

An integrative position is held by Boye (2012), who observes that ‘both notions (epistemic necessity and inferential justification) need to be invoked’ in order to account for the meaning and use of EM must and notes this is also the case of Spanish deber(de) (p. 24). The above caveats on the categorization of must as an evidential marker warrant our position in conceiving must as an EM, with evidential nuances derived from its conclusional force and including it in a subcategory of epistemicity distinct from that of evidentiality (Boye, 2012).

The texts

This article presents a case study which focuses on the expression of epistemic stance in journalistic discourse, based on CESJD-JMA. The texts were collected in a time period ranging from 1999 to 2010 from the sections on news reports and from the comment section (opinion columns and leading articles) of four quality papers equally differing in ideological orientation in both languages: The Guardian and The Times for English and ABC and El País for Spanish. Table 1 shows the number of words per genre.

OPEN IN VIEWER

Table 1.

Corpus of opinion columns, leading articles and news reports in English and Spanish (number of words).

English	Opinion	Leading	News	Total
1999–2010	NW	NW	NW	NW
English
The Guardian	55,878	56,498	61,187	173,563
The Times	55,195	55,027	60,255	170,477
Total	111,073	111,525	121,442	344,040
Spanish
ABC	51,687	51,416	59,623	162,726
El País	54,136	56,150	61,282	171,569
Total	105,823	107,566	120,905	334,294

NW: number of words.

Corpus study: Research questions and hypothesis

The corpus study explores the following issues regarding the presence and patterning of a selected set of stance markers in each sub-corpus:

It is hypothesized that there will be variation across genres due to the distinct goals and roles of writers of editorials or opinion columns, in contrast to news reports. Variation will also be found across languages due to possible intercultural differences in discourse practices in the argumentative and persuasive style or in the reporting style.

Research objectives and procedure

The subcategories of epistemic stance identified in this article and the set of expressions chosen are as follows. The focus is on those expressions which are judged to be more prototypical for each subdomain (introspective, top-down) and quantitatively more relevant in the texts (text-driven, bottom-up):

Multifunctional evidential expressions in journalistic discourse

Examples of multifunctionality in our comparable corpus are observed for a set of expressions in the two languages. As mentioned above, within the domain of evidentiality, the link between direct justification and indirect justification is evinced in both English and Spanish in the case of predicates invoking experiential evidence (see, se ve), where there is an extension of meaning from direct perceptual to IIE. Evidential values are also subject to extensions within the subdomain of indirect justification or indirect evidence. Both indirect-inferential and indirect-reportative functions are found for the following translation equivalent expressions: (a) English: apparently, appear, seem; (b) Spanish: aparentemente, al parecer, parece.

The verbs appear and seem or parecer attested in the corpus basically function with an indirect-inferential meaning (see examples (1), (3), (5) and (6)), although some instances are found which have developed with an indirect-reportative meaning. This phenomenon is common to other Germanic and Romance languages, as Wiemer (2010) notes, who makes the point that the 3SG form of the present indicative of ‘SEEM-verbs highlights a reportive function only when it loses its dependency relations with a host sentence, i.e., when it becomes petrified as a particle viz. parenthetical’ (p. 105). However, as the following examples show, SEEM-verbs in both languages are regularly found in our corpus, which do not comply with the above-stated condition but which have a reportative value:

(19) IRE: David Kelly did not live or belong in this world. He was a quiet, intensely focused microbiologist who had become an internationally respected expert in biological warfare, travelling frequently to Iraq as part of the UN weapons inspection programme. It appears <IRE> that, on an infrequent basis, he gave journalists private briefings, but in all other respects he lived a life as far removed as imaginable from the Westminster game. (ELG)

(20) IRE: There is some suspicion that things went wrong even before the landing at the South Pole itself: the two probes, Scott and Amundsen, designed to detach themselves, hurtle to the ground at 400 miles an hour and look for ice beneath the surface, went out of the tent and have been away for some time without any further word. It now seems <IRE> that not everything was fully tested before the mission started. But this is surely what comes of cutting costs. (ELT)

Spanish ‘parece’ shows an interesting parallelism with English ‘seem’ in its development of meaning, from the indirect-inferential (perceptual- or conceptual-based), and indirect-inferential (report-based), to the indirect-reportative meaning:

(21) IIE: En unos tiempos como los actuales en los que, al menos en la superficie de los acontecimientos, parece <IIE> idolatrarse el materialismo, siempre defendió la vida del espíritu, por supuesto compatible con la condición corporal del hombre, que también … (SLA)

 (In times like the present in which, at least on the surface of events, materialism seems to be idolized, he always defended the life of the spirit, …)

(22) IIE (+Report): A juzgar por sus declaraciones, parece <IIE> que Obama, prudente y comprometido con un sentido ético de la política, está dispuesto a gobernar desde el estricto respeto a la Constitución, las leyes y los derechos ciudadanos, … (SLP)

 (Judging by his declarations, it seems that Obama, prudent and committed to an ethical sense of justice, is ready to govern with a strict respect for the Constitution, the law and the civil rights)

(23) IRE: Esta respuesta era la que ETA pretendía de ellos al asesinar al ertzaina con tan brutal ensañamiento (le dispararon casi 30 tiros). Inmediatamente, se ha vuelto al discurso del ‘uno de los nuestros’, de tal manera que la muerte del concejal de UPN José Javier Múgica parece <IRE> ‘culpa del conflicto’, mientras que la de Uribe responde a un ‘conflicto interno’. (SLA)

 (… There has immediately been a return to the discourse of ‘one of our own’, in such a way that the death of the city councillor for UPN José Javier Múgica is supposedly ‘the fault of the conflict’, while that of Uribe is due to an ‘internal conflict’.

The parenthetical particle ‘según parece’ would be an instance where the meaning has become practically restricted to the reportative value:

(24) IRE: Se dice que sólo con el nombre de «Bar Las Vegas» hay más de mil distribuidos en nuestro territorio nacional. Aquí somos, según parece <IRE>, más vinícolas que alcohólicos, y nos gusta gastarnos el dinero en restaurantes y tertulias. (SOA)

 (… Here we are, apparently, more lovers of wine than alcoholics, and we like to spend our money in restaurants and talking in social gatherings.)

In the case of adverbial expressions, there seems to be a division of labour between aparentemente and al parecer in Spanish. Although both indirect-inferential and indirect-reportative meanings are found for these expressions, aparentemente tends to appear with an inferential meaning, whereas the particle al parecer is more common with the reportative meaning and also tends to appear as a parenthetical. Parenthetical uses of ‘al parecer’ with an indirect-inferential meaning, however, are also found in our corpus:

(25) IIE: Resulta que el amor es ciego y aparentemente <IIE> no le ha dejado ver al señor Rudolf Scharping (inepto jefe del SPD entre 1993 y 1995) que sus continuos desplazamientos en avión oficial para acudir a los brazos de la condesa Pilati en su villa de Mallorca no eran un comportamiento sensato. (SOA)

 (It turns out that love is blind and apparently has not allowed Mr Rudolf Scharping […] to see that his continuous trips in an official plane to fly to the arms of Countess Pilati in her villa in Mallorca did not constitute sensible behaviour.)

(26) IRE: Aunque muchos de los casos aún no han sido diagnosticados como legionelosis, el brote, al parecer <IRE> originado en el centro de la ciudad, puede haber afectado a más de cien personas. (SLA)

 (Though many of the cases have still not been diagnosed as legionnaire’s disease, the outbreak, apparently originating in the city centre, may have affected more than a hundred people.)

(27) IIE: Las relaciones internacionales de la era global se siguen jugando, al parecer <IIE>, en la química de las impresiones y los afectos. A pesar de la facilidad para establecer contactos interoceánicos instantáneos, nada como mirarse en la pupila sin intermediarios electrónicos, nada como olerse la fragancia varonil mezclándose con las fibras del tejido y los humores corporales, nada como rozarse la piel, pulsar el timbre de voz, hacer un comentario jocoso para ver cómo respira el otro. (SOA)

 (International relations in the global era are still played, it seems, in the chemistry of impressions and affect. In spite of the ease in establishing instant interoceanic contact, nothing like looking into each other’s eyes without electronic intermediaries, nothing like smelling one another’s manly fragrance …)

English apparently, however, shows a relatively equal proportion of both inferential and reportative uses:

(28) IIE: Now turn to Europe. A section of the Spanish electorate has apparently <IIE> decided that the appropriate response to the murder of 200 of their fellow citizens is to ask what their Government might have done to provoke this outrage. This is a stance that smacks of weakness, not defiance. (EOT)

(29) IRE: According to <IRE> intelligence briefings, his was the hand that cut Pearl’s throat, a murder for which the British-born militant Omar Saeed Sheikh has already been sentenced to death in Pakistan. The story of Sheikh’s ‘arrest’ was also clouded with inconsistencies. He had been involved in a string of terrorist acts and had served time in an Indian jail. He was released in a hostage swap, returned to Pakistan and apparently <IRE> continued his activities without undue interference from Pakistani intelligence. (EOG)

These examples point to a substantial amount of variation in the use of these adverbials and their extensions of meaning. Conclusions regarding the multifunctionality of these expressions are perforce tentative due to the limited number of examples found in the journalistic corpus.

Categories of epistemic stance in journalistic discourse

Table 2 shows the categories of epistemic stance and the frequency of occurrence of the expressions chosen for each language (raw numbers and normalized frequency in a ratio per 10,000 words), in the three genres of journalistic discourse.

OPEN IN VIEWER

Table 2.

Categories of epistemic stance in journalistic discourse in English and Spanish.

Journalistic genres
English	Opinion 111,073		Leading 111,525		News 121,442		Total 344,040
	N	R	N	R	N	R	N	R
IIE	90	8.10	89	7.98	45	3.71	224	6.51
IRE	48	4.32	26	2.33	81	6.66	155	4.50
EM	383	34.48	343	30.75	117	9.63	843	24.50
CGA	84	7.56	29	2.60	17	1.39	130	3.78
Total	605	54.46	487	43.67	260	21.41	1352	39.29
Spanish	Opinion 105,823		Leading 107,566		News 120,905		Total 334,294
	N	R	N	R	N	R	N	R
IIE	82	7.75	88	8.18	35	2.89	205	6.13
IRE	33	3.12	30	2.79	139	11.49	202	6.04
EM	168	15.87	142	13.20	73	6.04	383	11.45
CGA	97	9.17	39	3.63	2	0.17	138	4.12
Total	380	35.91	299	27.79	249	20.59	928	27.75

IIE: indirect-inferential evidence; IRE: indirect-reportative evidence; EM: epistemic modal; CGA: cognitive attitude.

Discussion of results

One of the striking features in the comparison of results across languages is that of the total figures. ² We observe that English by far surpasses the frequency of stance expressions in Spanish, which seems to point to a perceived need in English to provide both ‘evidential justification’ and, above all, ‘epistemic support’ for their claims to ‘knowledge’ or assertions. The normalized frequency for epistemic modality in English more than doubles the figure for Spanish. EMs found in the English corpus are for the most part expressions of partial support or uncertainty. This reflects a more tentative style in argumentative discourse strategies in English, involving a lesser degree of commitment to their assertions. As Bybee et al. (1994) observe concerning epistemicity and truth, ‘markers of epistemic modality indicate something about less than a total commitment by the speaker to the truth of the proposition’ (p. 179). This tendency is also reflected in the second preferred stance type in English, indirect-inferential justification. The characterizing features of these genres in English journalistic discourse would seem to be a lack of commitment in terms of justification and support and a marked preference for expressions indexing implicit subjectivity, such as EMs, and also expressions such as inferential evidentials which evoke a virtual or generalized conceptualizer, thus obfuscating personal responsibility and accountability for their assertions.

The pattern of distribution of stance categories in Spanish is very similar to that of English, although with some differences in the higher presence of reportative evidential expressions, especially in news reports, and in the category of CGA, where higher results are found in Spanish in both opinion and leading articles. This seems to point to a certain tendency to explicitly invoke the speaker as conceptualizer and assume personal responsibility for their assertions in Spanish argumentative discourse.

With regard to the similarities or differences across genres, there is a marked contrast between news reports and opinion and leading articles. Both opinion and leading articles have a similar pattern of distribution of the subcategories of epistemic stance, with the exception CGA expressions, which are considerably more frequent in opinion columns in both languages. This feature is clearly genre-related and coherent with the felt need to explicitly mark the presence of the speaker/conceptualizer in the discourse of opinion. In contrast, news reports are characterized in both languages by far lower frequencies in CGA expressions and in the use of epistemic modality and evidentials with an indirect-inferential value. In comparison with the other genres, we also find a genre-related feature in both languages in the higher frequencies for IREs.