Evidentiality and the verbal expression of belief and hearsay Elena Martínez Caro Universidad Complutense de Madrid Abstract: One of the areas of discussion within the study of evidentiality refers to how the speaker obtains knowledge, i.e. to the source of his/her information. Chafe (1986) distinguishes four ‘modes of knowing’ or different ways in which knowledge is acquired: belief, induction, hearsay and deduction. The present contribution investigates the correlation between evidentiality and the use of verbs denoting two of these forms of indirect evidence: knowledge steming from belief or opinion and knowledge having been acquired through language. This study specifically focuses on the use of verbs denoting mental cognitive and verbal processes as evidential markers in a corpus of newspaper discourse comprising press editorials and news reports in English and Spanish. Keywords: Evidentiality; Belief; Hearsay; Newspaper discourse; English; Spanish 1. Introduction Evidentiality is a semantic notion related to the kind of evidence a person has for making factual claims and how this is expressed in language (Anderson 1986:273). The notion originated with the early Americanists who studied forms with evidential meaning in the languages of the New World, but had little attention until a couple of decades ago when these forms began to be reported in other languages (Willet 1988:51). 1 There is no consensus as to the notional boundaries of evidentiality. Some authors adopt a narrow sense by which evidentiality centers around the sources of information or sources of knowledge behind assertions (as defined by Dendale and Tasmowski 2001:340). Others take a broader view by which evidentiality involves the linguistic expression of attitudes towards knowledge (as in Chafe 1986:262). In this second view, references to sources of information have been linked closely to references to certainty or reliability of knowledge (i.e. epistemic modality), because the linguistic markers encoding these two semantic domains are often the same (Dendale and Tasmowski 2001:340). The notion of evidentiality used here is closer to The first major step taken toward a deeper knowledge in this area was the Evidentials Symposium at Berkeley 1 in 1981, and the proceedings volume that emerged from it (Chafe and Nichols 1986). this second interpretation, although special focus is on the source of the information underlying the speaker or writer’s proposition. As Willett (1988:55) states, a common thread in the different views of evidentiality is that it “is the linguistic means of indicating how the speaker obtained the information on which s/he bases an assertion”. The strength of a speaker’s commitment to an assumption depends not only on the amount of evidence s/he has for his/her truth, but also on the type of evidence (Blakemore 1994:1183). Chafe (1986:263) distinguishes four modes of knowing or ways in which knowledge is acquired: belief, induction, hearsay and deduction. My aim here is to investigate the verbal expression of evidentiality focusing on the modes of belief and hearsay in English and Spanish newspaper discourse. 1.1. The linguistic expression of evidentiality Evidentials are expressed differently in natural languages. The set of grammatical devices used to indicate the type of evidence the speaker has for what s/he says includes a whole range of phenomena –syntactic, morphological, lexical and prosodic. Few languages mark evidentiality as a separate grammatical category; evidentiality is not grammaticalised in the languages we are most familiar with (Lazard 2001:366). English and other West European languages seem to primarily mark evidential contrasts lexically, usually as extended meanings or secondary usages (cf. Willett 1988:64). Evidential markers (and other epistemic stance markers, cf. Biber et al. 1999:972) are represented in English by various syntactic categories among which we find adverbs (probably, perhaps), adjectives (‘it is probable/certain), nouns (‘it is a possibility/certainty), verbs (I think/believe, it is said, it seems), modal auxiliaries (might, must, could) and parenthetical clauses or comment clauses such as I think. In this study I concentrate on the category of verbs expressing evidentiality in English and Spanish. These verbs occur as matrix predicates of complement clauses (1a-b) or as parenthetical clauses (examples (1c) and (2a-c)), sometimes preceded by as or como: (1) Verbs expressing belief: a. desde Platón sabemos que los dirigentes democráticamente elegidos no tienen por qué ser los mejores. (SEp31sm) 2 b. Mr Blair said: "I do believe that people can be persuaded that they have to pay for good public services... " (ENt06jm) c. This will, the CIA thinks, get worse. (ENg05spa) (2) Verbs expressing hearsay: a. Jury trial, it was said, was at the root of British criminal justice. (EEg02jm) b. «Lamentamos haber actuado tan tarde», ha dicho Schüssel, quien a la postre ha visto afeada su iniciativa por el líder de la comunidad israelí. (SNa04emc) c. Como bien dijo ayer el presidente Lagos, hay que respetar y acatar las decisiones de la Justicia. (SEa02jm) 1.2. Methodology All the examples considered in this paper come from a comparable corpus of news reports and press editorials reporting political events and various social issues. The texts were collected from liberal and conservative newspapers in English and Spanish, each group of texts in each language (e.g. editorials in English) amounting to a volume of circa 60,000 words. The examples contain information, in brackets, about the subcorpus to which they belong –the language, type of newspaper text, and newspaper— besides the number given to the text and the initials of the text analyser. 3 In order to identify evidential verbal markers of belief and hearsay, an electronic search of the tokens was carried out. The concordances which resulted were then revised to check their evidential value and whether they were verbs. Occurrences such as claim or doubt as See details of the notation conventions used in the examples in the following subsection (1.2.).2 The abbreviations used are (in order of occurrence in the examples): for language, E (English) and S (Spanish); 3 for text type, N (news reports) and E (editorials); for newspaper, t (The Times), g (The Guardian), a (Abc) and p (El País). nouns were, consequently, rejected, as well as instances of non-evidential homonymous senses of some of the verbs considered. I have concentrated on a restricted number of verbs per category in each language, bearing in mind that there could be other verbs expressing these two evidential meanings, although with an apparently lower occurrence. The majority of verbs were selected following previous studies on evidentiality discussing the expression of belief and hearsay (especially those on English) and given their basic meaning associated with these two domains. 2. The verbal expression of belief or opinion In this section I consider verbs showing that the writer’s knowledge is based on belief or opinion. These are verbs such as think, believe, suppose in English, and their counterparts 4 pensar, creer, suponer, considerar in Spanish. See the following examples: 5 (3) a. Phil Willis, the Liberal Democrat education spokesman, told The Guardian: "I think there is a real danger here of educational apartheid... " (ENg05jn) b. no one really believes that his views on homosexuality or, for that matter, on race and immigration have really changed. (EEg10sm) c. But it is a quite different matter to suppose that the British, or anyone else, can create a united Ireland. (EEg08sm) (4) a. Hoy somos muchos los que pensamos que, por mal que estén las cosas, un endereza- miento es posible. (SEp11jm) b. José Luis Rodríguez Zapatero cree que la debacle de los socialistas en Francia es "coyuntural" y que "recuperarán mucho terreno" en las elecciones legislativas del próximo junio. (SNa02jn) c. El censo aprobado el viernes por el Comité Federal del PSOE reduce la militancia real a 210.000, casi la mitad de lo que se suponía. (SNp09jn) d. Hay quien considera que estos problemas justifican la rebelión antiglobalizadora; muchos han considerado que las manifestaciones eran más democráticas que los gobernantes contra los que se rebelaban. (SEp31sm) I guess is often considered in the literature (cf. e.g. Chafe 1986), but is almost exclusively an American English 4 expression that occurs mostly in conversation (Biber et al. 1999:983). Considerar in Spanish has been here included with the sense of creer (believe or think), occurring as a matrix 5 predicate of a nominal subordinate clause as direct object expressing an assertion (cf. Delbecque and Lamiroy 1999:1997, 2008), as in example (4d). Consider in English, however, cannot be used with this sense as an assertive verb taking a complement clause (cf. Aijmer 1980:ch. 5). Examples of considerar in Spanish occurring with a NP direct object and an object predicative complement have been disregarded (as in “Consideran de vital importancia la entrega temporal de terroristas” (SNp06jn)). These verbs with an evidential meaning may be seen to have a double function. They may express the source of the writer’s knowledge or assertion, especially when used with a third person subject and referring to a particular person’s (or persons’) beliefs as evidence (cf. (5a-b)), or more specifically refer to the degree of reliability or certainty of the writer’s knowledge (as in (5c-d) where the writer relies on his/her personal opinion or knowledge): (5) a. One of those present at yesterday's meetings said Mr Mandelson believed Mr Blair's press secretary, Alastair Campbell, had been at fault. (ENg02jm) b. Para erradicar el terrorismo, así como para construir la paz, los Quince piensan que Israel necesita la Autoridad Palestina y a su presidente, Yasir Arafat, como interlocutor de la negociación. (SNp13mb) c. Lo decisivo, creo yo, el ámbito en el que España sigue manteniéndose a cierta distancia del resto de Europa, es el de la educación y el conocimiento. (SEp30sm) d. Tactical voting may see even more Tories culled than the polls are predicting, although personally I think this is being overstated. (EEg04sm) These two functions may, however, overlap and are sometimes difficult to separate. Concerning the second of these two functions, the verbs of belief and knowledge considered in this section can be seen as included within a scale of certainty of knowledge. Following this scale or continuum, two further verbs have been considered: know (saber in Spanish) in one of the extremes of the continuum, implying absolute/full certainty, and doubt (dudar in Spanish) closer to the other end (expressing very little certainty). Interestingly, however, the examples with doubt in English and dudar in Spanish are often affected by negation so that the evidential meaning is then similar to the one we get with know (as in (7a)): (6) a. It is well known that we have far more divorcees and far more teenage mothers than any other country in western Europe, but I am still puzzled why this is so. (EEg06jm) b. En este país todos sabemos qué es lo que tenemos que hacer para librarnos de la amenaza de ETA (SNp02jm) (7) a. And when the fears of radiation leaks do finally sail off into the Atlantic sunset, no one here doubts that Madrid, having also taken this episode on the chin for London, will also be exacting some price from Tony Blair. (ENg04spa) b. la mayoría de las explotaciones ganaderas son mixtas -vacas junto a otros animales destinados al consumo- y la Administración duda si los ganaderos han separado el pienso correctamente. (SNp01jm) With the predicates of belief and opinion shown in examples (3) and (4), there is always some degree of uncertainty involved, with suppose (suponer) ranking lower (i.e. implying ‘less certainty’) than think or believe (creer/pensar). All the verbs mentioned can be grouped as 6 predicates expressing epistemic attitudes (as in e.g. Aijmer 1980:37-38 and Biber et al. 1999:972-974), comprising in turn two subfields: one corresponding to knowledge and another to belief. Verbs such as think, believe, suppose, etc., in which the writer is less than “wholeheartedly committed to the truth of the proposition (and by implication that [his] grounds are not strong for him to use the unqualified assertion)” (Aijmer 1980:25), make it possible for the writer to use them as hedges or mitigating devices: Because the mental-state predicates are inherently subjective, they are frequently used as mitigating or hedging devices. (...) In such uses, it is quite obvious that speakers are absolutely certain about or convinced of what they are saying, but by using the mental state predicate they suggest that they are voicing a tentative and personal opinion which may be wrong, thus ‘officially’ leaving room for another opinion or for a reaction on the part of the hearer. (Nuyts 2001:391) In the examples of mental-state verbs expressing belief and knowledge, several grammatical forms have been found in relation to the subject accompanying the verb (or its omission in Spanish) and the choice of active/passive voice. These different grammatical forms can be related to the extent to which the attitude or stance expressed is attributed to the speaker/ writer. Following Biber et al. (1999:976) three degrees can be distinguished within evidential forms: explicit attribution of attitude/evidentiality to speaker/writer, implicit attribution (i.e. attitude/evidentiality may be expressed without overtly identifying the speaker/writer) and, finally, ambiguous (or, we could say, impersonal) attribution of attitude/evidentiality. In the first group we can include instances with a first-person (both singular and plural) pronoun as As Aijmer (1980:18) notes, whereas know is a factive and assertive predicate, predicates related to ‘to 6 believe’are non-factive and assertive. subject (or its omission and the explicit verbal inflection in Spanish), which make the attribution overt: (8) a. Mr Blair said: "I do believe that people can be persuaded that they have to pay for good public services..." (ENt06jm) b. "We know that more than 250 reserve soldiers actually refused," he said. He added: "I think now it is more serious than ever before..." (ENgspa) (9) a. Lo decisivo, creo yo, el ámbito en el que España sigue manteniéndose a cierta distancia del resto de Europa, es el de la educación y el conocimiento. b. Uno creía que los españoles éramos grandes bebedores, pero la OMS, Organización Mundial de la Salud, nos ha demostrado lo contrario. (SEa30sm) In (9b), the writer uses the indefinite pronoun uno as a substitute for yo, giving the sentence a more impersonal tone. Other examples with third-person pronouns or full noun phrases can be also included in this group, showing that the attitude is attributed to a third person rather than to the writer: (10) a. Mr Arafat may not give these men of violence express orders; he has, at long last, had the grace (or sense of self-preservation) to condemn this particular attack. But he knows who they are (EEt04sm) b.Afghan delegates, who will submit an alternative needs estimate, believe the Tokyo conference is their best chance of securing funds before the attention of the world moves elsewhere. (ENg16mb) (11) a. La presidencia española, según supo ABC, sigue pensando que la Unión Europea no puede dejar pasar su responsabilidad en relación con Oriente Medio y quedarse de brazos cruzados mientras la situación sigue deteriorándose (SNa09mb) b. Teo Uriarte cree que la presión terrorista de ETA ha hecho regresar la pesada atmósfera de los estados de excepción. (SNp02jm) c. Daniel Ben Simón, politólogo y editorialista del diario «Haaretz», considera, en efecto, que la sociedad israelí es "una sociedad tribal, en la que cada sector, cada tribu, lucha por sus intereses y prioridades particulares en detrimento de la colectividad nacional". (SNa02emc) With non-finite forms of these verbs, the attribution of attitude is often not overt but may be implicitly attributed to the writer with a possible implication of the reader as well (cf. 12a-b, 13b-c), thus showing some degree of impersonality, or to what a third party thinks (13a): (12) a. It is time to think about what we want in Afghanistan as the outcome of this war. (EEg08lh) b. It is barely possible to believe that less than three months ago, the main preoccupation of the then chief executive and his finance director, John Mayo, was to persuade shareholders to allow staff share options to be rebased as reason began to return to investors and the share price tumbled. (EEt12sm) (13) a. Sin embargo, se excluye de la prohibición el rabo, al creer que no existe riesgo por la ausencia de ganglios. (SNa02spa) b. Desde una nueva demarcación judicial con la colaboración activa de las autonomías hasta un sistema más estricto de responsabilidad de los jueces y del Estado por errores y negligencias, pasando por el reconocimiento expreso de 'una carta de derechos' de los usuarios de la justicia, todo hace pensar que el protagonista de la reforma es el ciudadano, como subrayó el representante del PSOE en la presentación del acuerdo. (SEp03lh) c. Con el cúmulo de oficiosidades perpetradas desde el entorno de La Moncloa (es lógico suponer que a impulsos del actual inquilino) para ningunear a González en sus presencias públicas dentro y fuera de España, se podrían escribir varios tomos de letra apretada. (SEp03sm) Finally, Biber et al. (1999:977-978) mention examples with “short passives” as ambiguous as to whether they mark the writer’s attitude or that of some other party. In our view, however, they are best regarded as having an impersonal generic interpretation, potentially representing all humanity (Kitagawa and Lehrer 1990:742) and not overtly attributing the proposition to a specific person. See the following examples from our data: (14) a. It is well known that we have far more divorcees and far more teenage mothers than any other country in western Europe, but I am still puzzled why this is so. (EEg06jm) b. As for the mass of teenage pregnancies, we are generally thought to be shyer and more inhibited about sex than our fellow Europeans (EEg06jm) c. Iraq, a likely focus of future US action, is believed to have built a series of deep bunkers to house biological and chemical weapons as well as command centres (ENt14jm) (15) a. El censo aprobado el viernes por el Comité Federal del PSOE reduce la militancia real a 210.000, casi la mitad de lo que se suponía. (SNp09jn) b. Los dos finalistas que salgan hoy del voto emitido por los parlamentarios conservadores deberán someterse al sufragio de las bases. Es bien sabido que entre ellas, Clarke es con mucho el más popular de los posibles candidatos. (SEa06sm) c. Dicho esto, se considera, no obstante, que es mejor no precipitarse a la hora de tomar una posición y reflexionar más sobre las aportaciones hechas por distintos países como Francia, Italia y Alemania. (SNa09mb) The use of this agentless passive avoids the need to mention the subject or agent of the evidential verb. “In news, this seems a deliberate strategy to avoid direct responsibility for the reported stance” (Biber et al. 1999:977). Some of these examples can be also considered to belong to the category of “knowledge derived from tradition or common knowledge” (Plungian 2001:352, Lazard 2001:365). Examples of evidential forms with an impersonal interpretation may evaluate or qualify the proposition they assert by means of adjectives or adverbs which generally refer to the degree of reliability of the information or may even evaluate the assertion itself (cf. 16). By using these qualifications, the writer authorises his/her readers “to put more or less faith in the assertions or conclusions, to bank on them, rely on them, to treat them as more or less trustworthy” (Toulmin 1958:90f, in Aijmer (1980:25-26). See examples (13c), (14a) and (15b) above and the following: (16) He is entitled to think that this has been the single most successful element of policy in the lifetime of this Parliament. (EEt11lh) (17) Se ha frenado el intento de algunas empresas de explotar líneas conjuntamente o sumar slots (...) en aeropuertos punteros y se ha permitido así la participación de un mayor número de compañías en el mercado. Pero en situaciones como las actuales no es disparatado pensar en una mayor flexibilidad en la ordenación del negocio. (SEp32lh) Other examples of predicates expressing belief or knowledge with an impersonal interpretation are those which have subjects expressed by personal pronouns used impersonally, indefinite pronouns such as everyone/no one or alguien/nadie or other NPs which are not fully specified by the writer of the news report or editorial. These subjects can also be used as a strategy to conceal the attribution, thus avoiding responsability, or leaving it unspecified or ambiguous. In examples with first person plural pronouns (or their omission in Spanish), there may be a “corporative reference” (Fernández Soriano and Táboas Baylín 1999:1729) involved in which the person(s) as subject assert(s) something, not in individual terms, but as member(s) and representative(s) of the group or institution they belong to (cf. Martínez Caro 2002:77): (18) She added: "We believe that this is a racist practice... " (ENg05jm) (19) a. "...Creemos que, en este año, hemos cambiado a Austria en la dirección necesaria". (SNa04emc) b. Pero la gente considera que, al fallar en un sentido o en otro, uno está involucrándose políticamente. (SNp06spa) 2.1. The expression of belief in English newspaper discourse Table 1 below shows the number of occurrences of mental-state predicates expressing belief or knowledge found in the English texts. Within each type of text, differences between conservative (Times) and liberal (Guardian) newspapers are shown: Table 1: The expression of BELIEF in English newspaper discourse As can be observed, there is a clear difference between the verbs know, think and believe and the other two verbs considered (suppose and doubt). Whereas the first group of verbs show occurrences of around 30% of the total number of verbs found in English (198), suppose and doubt represent only around 4% and 2%, respectively. The two text types considered, one expressing facts (news reports) and the other opinion (editorials), show no obvious difference in their use of these verbs, following the pattern described. The total number of mental-state predicates expressing belief or knowledge found in the English texts (198) represent 1.6 occurrences per thousand words. Considering the three most common verbs used and the types of subjects occurring with them, third person pronoun subjects show the highest frequency in all three of them, followed by first person ones. Among the latter, the use of a first person singular pronoun subject is particularly common with the verb think (I think). In the case of know, I know shows a similar NEWS REPORTS (62,766 w.) EDITORIALS (61,399 w.) TOTAL Guardian Times Total Guardian Times Total KNOW 15 19 34 15 19 34 68 THINK 13 13 26 18 17 35 61 BELIEVE 14 17 31 9 17 26 57 SUPPOSE 0 2 2 2 4 6 8 DOUBT 3 0 3 1 0 1 4 TOTAL 45 51 96 45 57 102 198 frequency of occurrence to we know, and with believe, the instances of first person (singular and plural) pronoun subjects show much lower figures compared to X believe(s). As for the use of the passive voice in all the verbs considered, 23 instances of passives were found, representing 11.56% of the total figure of belief and knowledge vebs. The higher proportion was found in news reports and, among the verbs, know, suppose and think occurred more frequently with the passive. 2.2. The expression of belief in Spanish newspaper discourse Table 2 shows the results related to the mental-state predicates expressing belief and knowledge found in the Spanish texts, both in the liberal (El País) and conservative (Abc) newspapers: Table 2: The expression of BELIEF in Spanish newspaper discourse The figures in Spanish differ in a number of aspects from the ones observed in English. First, whereas the number of verbs considered is similar in news reports to the figure in English in the same text type, the frequency in editorials is somewhat lower. Consequently, NEWS REPORTS (63,901 w.) EDITORIALS (61,196 w.) TOTAL El País Abc Total El País Abc Total SABER 7 15 22 10 15 25 47 PENSAR 11 1 12 10 4 14 26 CREER 15 16 31 9 5 14 45 SUPONER 5 8 13 6 2 8 21 CONSIDERAR 5 7 12 4 2 6 18 DUDAR 2 1 3 0 0 0 3 TOTAL 45 48 93 39 28 67 160 the total number of instances in the two text types in Spanish is also lower (160), representing 1.3 occurrences per thousand words. Second, in terms of the frequencies the specific verbs show, saber and creer occur most frequently in both text types, representing 29.4% and 28.1% of the total number of verbs found in Spanish, respectively, followed by pensar (16.3%), suponer (13.1%) and considerar (11.3%). Dudar is very infrequent, used with an occurrence of only 1.9%. As for the grammatical person these verbs agree with (reflected in the type of explicit subject and/or the verbal inflexion), the most common is the third person (both singular and plural, 40.6%), as in the English data. This mainly occurs with creer, saber and considerar, and more often in news reports. Verbs referring to a third person singular have the highest frequency of all (as in X cree/sabe/considera que...). Next in frequency (23.8%) we have the first person, either singular or plural, mainly occurring with creer and saber. The first person singular is used almost two times more and a great number of these instances appear in the expression (yo) creo que... The rest of the examples are verbs in a non-finite form (infinitive, or present or past participle) and passives. Non-finite verbs are either ambiguous in their attribution (having an impersonal interpretation) or they may implicitly identify the speaker/writer. Particularly common is the use of the past participle with suponer: supuesto/a/os/as. The passive represents 8.1% and it is mainly a passive with se (Se sabe/cree/supone/considera que...). 3. The verbal expression of hearsay In this section I consider examples which express knowledge having been acquired through language rather than direct experience. As Willett (1988:57) points out, the primary evidential parameter within natural language is that of direct evidence versus indirect evidence; that is, “whether the source of the speaker’s information is of a primary or a secondary nature”. Hearsay can be seen as one of the two types within indirect evidence: evidence via verbal report (the other being evidence upon which an inference is based; cf. Willett 1988:57). The verbs considered in this section include basic verbal processes such as say, tell, claim and report in English, and their counterparts decir, relatar and declarar in Spanish. Others were added like allege in English (and its Spanish counterpart alegar), with a clear evidential meaning, or indicar and asegurar in Spanish, which also appeared to a great extent in our texts. Some examples are the following: (20) a. "All history shows that if you allow disorder to take over from democracy then fascism can win," he said at the weekend. (EEg04jm) b. Terje Roed-Larsen, the UN's special coordinator for the Middle East peace process, t o l d t h e P a r i s c o n f e r e n c e : " T h e s i t u a t i o n w a s s o b a d t h a t a n y incident could have triggered the violence. It was a crisis in waiting." (EEg01lh) c. Some have tried to play down the significance of the abstentions, by claiming that people did not vote because the result was a foregone conclusion. (EEt05jm) d. Time magazine reported that intelligence officials believed bin Laden's taped broadcast on the first day of the airstrikes contained a coded warning to start more attacks. (ENt09spa) e. Conservatives were last night tabling parliamentary questions demanding to know the identity of the civil servants who are alleged to have given hostile press briefings against Ms Filkin during her turbulent three years as Standards Commissioner. (ENt05jn) (21) a. Tras los atentados del 11 de septiembre se dijo que entre las personas detenidas en Estados Unidos relacionadas con aquellos acontecimientos había tres españoles. (SNa08jm) b. Mario Onaindia relata la extorsión que representa la escolta para la familia y las relaciones sociales del protegido (SNp02jm) c. "La crisis de las vacas locas ha dado al traste con los cálculos previstos", declaró ayer la comisaria de Presupuestos, Michaele Schreyer. (SNp03spa) d. No está consolidado, se alega. (SEp02sm) e. Fuentes del CGPJ indicaron que previsiblemente el puesto de Jefe de Sección de la Escuela será para Ignacio Sancho Gargallo, magistrado del Juzgado de Primera Instancia 31 de Barcelona. (SNa10jn) f. "La tropa es un bien escaso", asegura el jefe del Estado Mayor de la Defensa, almirante Antonio Moreno Barberá, en la entrevista que hoy publicamos en páginas de Nacional. (SEa01sm) Whereas “‘ordinary’, non-evidential forms state the facts purely and simply, [e]vidential forms, on the other hand, point to the speaker’s becoming aware of the facts” (Lazard 2001:362, my italics). In the case of hearsay, for example, the evidential implies “as I have heard”. The operation may be seen to create a distance “between the speaker and his/her own discourse, or between the speaker as the person acquiring evidence and the person expressing it” (Lazard 2001:362) and has been thus characterised as ‘mediative’ evidence. Hearsay examples may refer to information that has been reported using the exact words uttered (direct speech) or a reproduction of them given by the speaker/writer (indirect speech). In the first case the quotation need not extent to the whole clause reported as in (22b): (22) Direct speech in English and Spanish: a. "The problem is the daily indignities and injustices, the checkpoints and the barriers that impede movement and stop people from going to work," he told reporters last June. (ENg03spa) b. He says it was "a policy of promises" and no real change to the position of Albanians living in Macedonia that prompted them to create the guerrilla army. (ENg02--) c. "It's as if it was always there," says one fulltime observer. (ENg04jn) d. "Aquéllos que pudiendo trabajar no quieren hacerlo, no pueden vivir a costa de los demás", dijo el presidente (SNa07jn) (23) Indirect speech in English and Spanish: a. However the Egyptian president, Hosni Mubarak, told Mr Sharon that he risked increasing violence if he did not stick to the understanding reached by the previous Israeli government. (ENg02emc) b. The prime minister is pushing social justice and enterprise to the top of his agenda, reports political correspondent Nicholas Watt. (ENg03jm) c. El Papa Juan Pablo II, en su mensaje anual sobre "El estado del mundo", dijo el sábado que la crisis de Oriente Medio corría el peligro de desbordarse (SNa03emc) The direct speech quotation often occupies the initial position in the clause, thus foregrounding the information conveyed by it and backgrounding the subject entity and predicate which often transmit known or given information, or at least considered to be less informative than the quotation itself (as e.g. in (22a)). The Spanish examples with initial quotations often invert subject and verb, the latter being “triggered” by the object to occupy the second position, thus resulting in an OVS pattern. This structure may occasionally be found also in English (as in 23b). Examples based on hearsay evidence can be grouped in two categories: the information may come from reported speech from an author, that may be more or less known or specific; or it may be based on common knowledge or tradition (Plungian 2001:352). In the first case, examples occur with subjects of the reporting verbs which range from identifiable, known authors, to less specific or identifiable and more ‘impersonal’ ones. A selection of these is shown in (24) and (25): (24) a. Daniel Serwer, a former US diplomat and now a member of the US Institute of Peace in Washington, says the financing "poses as much a risk to US soldiers in Kosovo as the Albanian gunmen crossing over the border into Macedonia". (ENg02emc) b. Time magazine reported that intelligence officials believed bin Laden's taped broadcast on the first day of the airstrikes contained a coded warning to start more attacks. (ENt09spa) c. "Our people are being very cautious," one palace official told the Guardian. (ENg01jn) d. Since the petrol crisis, the Dome and other similar issues, there has been a loss of trust in both Tony Blair and Gordon Brown. Forty nine per cent say they find the Prime Minister untrustworthy and 48 per cent say they find the Chancellor untrustworthy." (25) a. Un año después de verse bajo sospecha, vejada y abandonada por sus socios, Viena dice haber dado sopas con onda a sus críticos y cambiado para mejor. (SNa04emc) b. Fuentes del CGPJ indicaron que previsiblemente el puesto de Jefe de Sección de la Escuela será para Ignacio Sancho Gargallo, magistrado del Juzgado de Primera Instancia 31 de Barcelona. (SNa10jn) In this connection, a correlation can be pointed out between the strength of an assumption based on reported speech and the degree of confidence one has in the person having reported it (Blakemore 1994:1183) and, also, to a certain extent, on the degree of knowledge the writer has about the ‘reporter’. In cases in which the author of the attribution may be unknown or considered unimportant by the writer of the newspaper text, or s/he may want to avoid an explicit mention to it, we may find metonymic expressions referring to inanimate entities as subjects: (26) a. A new report says that since 1970, 250,000 Protestants in Northern Ireland have been forced out of the areas they lived in. (EEg08sm) b. In 1999, a judicial investigation claimed that "at least 48 people" were killed in the massacre. (ENg07emc) These show some similarity to metonymies taking as subject of the hearsay predicate countries, newspapers or institutions rather than the persons ruling these, as in “Viena” in (25a) or “Time magazine” in (24b). In the examples in (26), however, there is a higher degree of mystification of the agent or person responsible for the assertion. Hearsay predicates which do not make explicit reference to the attribution of the reported information may be classified within the category of knowledge derived from tradition or common knowledge, as with predicates of belief and opinion. These are mainly examples with passives and subjectless non-finite clauses, or containing an impersonal third person plural pronoun (or its omission in Spanish, cf. (28a)): (27) a. Displaced fishermen and farmers will, it is always claimed, find new and lucrative jobs through tourist development, but it is rare for anybody to ask them whether they would actually rather be waiters or cleaners, or carry on as they were. (EEt01jm) b. In the same article in The Times, the author wrote: "It is often said that having children is a personal choice and it is a decision for which, as parents, we are expected to take sole responsibility... " (EEt08sm) (28) a. Dicen que quien dejó de ser tu amigo es que en realidad no lo fue nunca. Ha venido a mi memoria ese viejo dicho -lleno de sabiduría popular... (SEa10jm) b. No obstante, no salgo de mi asombro, porque España es el país con más bares, tabernas y ventas del mundo. Se dice que sólo con el nombre de «Bar Las Vegas» hay más de mil distribuidos en nuestro territorio nacional. (SEa30sm) In these examples expressing common knowledge the use of adverbials with a universal or generic meaning may favour the impersonal interpretation (cf. Fernández Soriano and Táboas Baylín 1999:1736), as always and often in (27a-b). In other examples without overt attribution, the writer of the newspaper text may obscure the attribution to a specific entity, directing the assertion to a third party or simply leaving the subject responsible for reporting the information unspecified, thus avoiding responsibility for the assertion: (29) a. Defence counsel are said to be scratching around to find any legal grounds for appeal. (EEg02jm) b. A further 845 people have been reported missing by their families in the UK, raising fears that the final British death toll could rise further. (ENt12mb) c. My conversations tell me that the commitment to build and deploy is likely to be made: much more likely, at any rate, than not. (ENg05spa) d. There are reported to be a number of other Britons at the base, but US officials did not confirm this. (ENg15mb) e. Los rabinos ultraortodoxos han vendido su voto a Sharon por un apoyo temporal a la política de exención del servicio militar para los estudiantes de los seminarios talmúdicos, así como por ayudas económicas a sus redes de escuelas y hospitales, según se asegura en círculos políticos (SNp07emc) It seems or it appears (parece in Spanish) have been mentioned in the literature as having a possible hearsay interpretation (cf. e.g. Chafe 1986). See the following example in Spanish, where the expression según parece has the evidential value of tradition or common knowledge, with a similar interpretation to expressions such as todos dicen, se dice, se cree... These expressions have not been considered in our analysis: (30) Aquí somos, según parece, más vinícolas que alcohólicos, y nos gusta gastarnos el dinero en restaurantes y tertulias. (SEa30sm) 3.1. The expression of hearsay in English newspaper discourse See Table 3, which shows the results related to the predicates expressing hearsay in the English data, in the liberal and conservative newspapers: Table 3: The expression of HEARSAY in English newspaper discourse If we compare these figures with the ones shown in Table 1, related to the predicates expressing belief and opinion in English, the first striking difference is the much higher number of occurrences, especially in news reports where the figure is more than five times higher. Over the total number of words analysed in this language (124,165 words), we found 5.8 English hearsay predicates per thousand words. NEWS REPORTS (62,766 w.) EDITORIALS (61,399 w.) TOTAL Guardian Times Total Guardian Times Total SAY 181 275 456 63 49 112 568 TELL 23 30 53 12 9 21 74 CLAIM 11 14 25 10 10 20 45 ALLEGE 2 4 6 0 0 0 6 REPORT 7 9 16 4 1 5 21 TOTAL 224 332 556 89 69 158 714 The figures also differ considerably if we compare the two text types analysed, where the use of evidential verbal markers of hearsay in news reports is 3.5 higher than their use in editorials. This same tendency was also attested for Spanish (cf. 3.2). This may be motivated by the fact that in news reports the writer intends to distance himself somehow from what he is saying, often relying on what others have reported, in order to present his/her information as objective. Quotations are, therefore, extremely frequent in this type of texts (compared to editorials which are more personal). Among the set of verbs, say is extremely frequent with a percentage of almost 80% over the total, followed by tell (10.4%), claim (6.3%) and report (2.9%). Comparing news reports and editorials, say occurs somehow higher in news reports, whereas tell and, especially, claim show higher relative figures in editorials. Allege is infrequent in news reports and not used at all in editorials, occurring only in 0.8% of the total number of instances. Looking at the kind of attribution associated with these verbs, an overwhelming number of predicates attributes the proposition to a third person, especially singular with say (X says/ said...) in news reports (in 90.8% of the total). Third person plural subjects also occur, although much less frequently in comparison (13.6%). We have also found a few instances of first person subjects but often associated with a third party as well, when they occur in quotations of what a third person reports. Further, the attribution may be left unspecified when a passive or non-finite form of the verb is used. The passive is used 4.1% of the total and, comparing the different verbs and their frequency, it occurs more frequently with report (as in it is reported, X was reported), allege (X are alleged) which also occurs as an –ed participle, claim (as in it is claimed) and tell (it/X was told). 3.2. The expression of hearsay in Spanish newspaper discourse Finally, Table 4 shows the results regarding the verbal expression of hearsay in Spanish newspaper discourse: Table 4: The expression of HEARSAY in Spanish newspaper discourse Comparing the figures with the ones shown for English in Table 3, we can notice the comparable lower frequency of the Spanish figures, especially in editorials. Over the total number of words analysed in this language (125,097 words), we have found 2 evidential markers expressing hearsay in Spanish per thousand words, the great majority in news reports (1.7 per thousand words) vs. editorials. Among the specific predicates found, decir –as say in English and considered the most neutral or basic communicative predicate in Spanish— occurs most often (39.5% of the total figure), followed by asegurar (23.9%), comparatively much more frequent in news reports than in editorials, declarar (18.2%) and indicar (10.1%). Alegar and relatar show the lowest figures. In terms of attribution, the evidential predicates expressing hearsay in Spanish attribute the reported proposition to a third party almost exclusively, especially to a third person singular subject (77.7%), although third person plural subjects also occur (18.2%). A few NEWS REPORTS (63,901 w.) EDITORIALS (61,196 w.) TOTAL El País Abc Total El País Abc Total DECIR 42 43 85 7 12 19 104 RELATAR 2 1 3 0 0 0 3 DECLARAR 19 19 38 3 4 7 45 ALEGAR 5 3 8 2 1 3 11 INDICAR 9 15 24 1 0 1 25 ASEGURAR 33 24 57 0 2 2 59 TOTAL 110 105 215 13 19 32 247 occurrences of these verbs in the passive have also been found (3.6%), with decir, asegurar and alegar, all of them instances of se-passive: se dice/se ha dicho/se dijo/se decía, se asegura, se alega. 4. Conclusion In this paper, I have discussed verbal markers of evidentiality in English and Spanish looking at two types of evidence, belief and hearsay, within the discourse of fact and opinion in newspaper discourse. Blakemore (1994:1185) (following Blass and Itani-Kaufman) sees a point in common in the two types of evidence discussed here. In some uses, constructions such as It is said that often mean it is generally believed that. Aijmer (1980) points to the 7 same similarity of function of the two expressions, when the meaning is of common knowledge: “The fact that something is said by many or by an institution is evidence for believing it to be true” (Aijmer 1980:119). In a study on the correlation between impersonalisation and the use of impersonal pronouns and other non-specific NPs in English and Spanish news reports (Martínez Caro 2002:85), the use of a verb denoting a process typically associated with human beings, such as a verb of thinking or saying, especially in examples containing a first-person plural or third-person plural pronoun (or their omission in Spanish) was also reported as a common feature of a considerable number of the non-specific agent constructions discussed. Willett (1988:86) proposes a hypothesis describing the relation between the source of the information used in an assertion and the strength with which it is asserted, which also connects the two modes of knowing discussed in this paper. Concerning the source of a Blass reports that in a language such as Sissala, hearsay particles are found both in examples of reported 7 speech as well as in examples which involve verbs of belief and propositional attitude (think, believe, hope, etc.). speaker’s information, he considers the three major evidential parameters he distinguishes within natural language: direct or attested evidence (sensory evidence), and within indirect evidence, reported (evidence based on hearsay) and inferring (where he includes evidence based on belief.) Willett’s hypothesis states that, on a scale from most to least direct, attested 8 evidence is ranked as the most reliable source, inferring evidence as the least reliable, and reported evidence as somewhere in the middle. Thus, an assertion based on sensory evidence will be presented as “certain, or perhaps emphasized by the use of the appropriate language- specific devices”; if it is based on hearsay evidence, the assertion will be expressed as less certain; further, if it is based on inference, “the assertion will likely reflect even less certainty and more probability” (Willett 1988:87-88). References Aijmer, K. (1980) Evidence and the Declarative Sentence. (Stockholm Studies in English.) Stockholm. Anderson, L. (1986) Evidentials, paths of change, and mental gaps: typologically regular asymmetries”. In Chafe and Nichols (eds.), 273-312. Biber, D., S. Johansson, G. Leech, S. Conrad and E. Finegan (1999) Longman Grammar of Spoken and Written English. London: Longman. Blakemore, D. (1994) Evidence and modality. In J.M.Y. Asher (ed.) The Encyclopedia of Language and Linguistics. Oxford: Pergamon. 1183-1186. Chafe, W. (1986) Evidentiality in English conversation and academic writing. In Chafe and Nichols (eds.), 261-272. Chafe, W. and J. Nichols (eds.) (1986) Evidentiality: The Linguistic Coding of Epistemology. Norwood, NJ: Ablex. According to Willett (1988:90, n.10), “what Chafe (1986) calls ‘belief’ appears to correspond to an inferential 8 based on unspecified evidence”. Delbecque, N. and B. Lamiroy (1999) La subordinación sustantiva: Las subordinadas enunciativas en los complementos verbales. In Bosque, I. and V. Demonte (eds.) Gramática descriptiva de la lengua española. Vol. 2. Madrid: Espasa-Calpe. 1965-2081. Dendale, P. and L. Tasmowski (2001) Introduction: Evidentiality and related notions. Journal of Pragmatics 33: 339-348. Fernández-Soriano, O. and S. Táboas Baylín (1999) Construcciones impersonales no reflejas. In Bosque, I. and V. Demonte (eds.) Gramática descriptiva de la lengua española. Vol. 2. Madrid: Espasa-Calpe. 1723-1778. Kitagawa, C. and A. Lehrer (1990) Impersonal uses of personal pronouns. Journal of Pragmatics 14: 739-759. Lazard, G. (2001) On the grammaticalization of evidentiality. Journal of Pragmatics 33: 359-367. Martínez Caro, E. (2002) Impersonalisation and reference in English and Spanish: Evidence from newspaper discourse. In J. I. Marín Arrese (ed.) Conceptualization of Events in Newspaper Discourse: Mystification of Agency and Dedree of Implication in News Reports. Proyectos Complutense 2000, Universidad Complutense de Madrid. 71-88. Nuyts, J. (2001) Subjectivity as an evidential dimension in epistemic modal expressions. Journal of Pragmatics 33: 383-400. Plungian, V.A. (2001) The place of evidentiality within the universal grammatical space. Journal of Pragmatics 33: 349-357. Toulmin, S. (1958) The Uses of Argument. Cambridge. Willett, T. (1988) A cross-linguistic survey of the grammaticization of evidentiality. Studies in Language 12(1): 51-97.