<<

Perspectives in Reported David Y. Oshima

Perspectives in Reported Discourse

David Y. Oshima

Stanford University June 3rd, 2005 Dissertation Proposal Talk

Stanford University Page 1 Perspectives in Reported Discourse David Y. Oshima

Background

- The notion of perspective (and its subvarieties) plays an important role in many aspects of natural languages, such as:

• construction alternations, (syntax)

• deixis and ()

• functional sentence perspective (semantics- interface)

• narrative styles, rhetorical structures (discourse)

Stanford University Page 2 Perspectives in Reported Discourse David Y. Oshima

- Various notions subsumed by or closely related to linguistic perspective have been discussed in the literature: spatial/temporal/social deixis, empathy, topicality, attention , figure- ground, subjectivity, ...

- The exact cognitive/ontological natures of these notions, and the in- terrelations, similarities, and differences among them, are far from well- understood.

Stanford University Page 3 Perspectives in Reported Discourse David Y. Oshima

- Reported discourse is an especially important domain in this light. When an agent reports another agent’s utterance (or belief, etc.), he can choose from, or otherwise mix, two standpoints from which the utterance is rep- resented: that of the reporter-agent and that of the reported-agent.

- Dynamic interactions, fusion, and clash of two perspectives within re- ported discourse provide us with important clues as to the nature of many perspectival features in natural languages.

Stanford University Page 4 Perspectives in Reported Discourse David Y. Oshima

The aim and organization

- The main goal of this thesis is to develop semantic analyses of attitude reports that capture various effects of perspective-taking.

- First, I establish the distinction between the direct and indirect modes of reported discourse. A report in the direct mode describes a relation between an individual and a linguistic object (linguistic representation), while a report in the indirect mode describes a relation between an indi- vidual and a semantic object (, etc.) (Partee 1973).

Stanford University Page 5 Perspectives in Reported Discourse David Y. Oshima

- Then, I demonstrate that the indirect mode of attitude reports can be divided into subtypes, depending on the extent to which it maintains the “original” perspective.

- Certain indirect attitude reports are more “direct speech-like” than oth- ers, in the sense that they contain perspectival features that are anchored to the secondary agent (whose utterance, belief, etc. is reported). I take up four types of “perspective shift” phenomena:

1. the de re/de dicto opposition (the analytic perspective shift)

2. the de se/non-de se opposition (the logophoric perspective shift)

3. the deictic perspective shift

4. the empathic perspective shift

Stanford University Page 6 Perspectives in Reported Discourse David Y. Oshima

- Finally, I examine factors that affect/condition the speaker’s choice of perspective in discourse reporting:

1. vividness of a report (Tannen 1985)

2. the hierarchy of attitude predicates (SAY > BELIEVE > KNOW > SEE; Culy 1994)

3. implicational relations between the subtypes of perspective

Stanford University Page 7 Perspectives in Reported Discourse David Y. Oshima

The two modes of reported discourse

- It is traditionally well-known that there are two kinds of reported dis- course – direct and indirect.

(1) a. John said (to me): “I need to talk to you”. b. John said (to me) that he needed to talk to me.

Stanford University Page 8 Perspectives in Reported Discourse David Y. Oshima

- The two modes differ in various respects: • verbatimness • syntactic marking (e.g. mood, complementizer) • deictic adjustment • expressives (e.g. “gee”, “uh”, “what a story!”) • syntactic well-formedness (John said: “ate, John, peanuts” vs. *John said that ate John peanuts) • syntactic opacity (*What did John say: “I want to eat ”?), ...

- What is the definitive feature that distinguishes the two modes?

Stanford University Page 9 Perspectives in Reported Discourse David Y. Oshima

Direct reports: relations between agents and linguistic objects Indirect reports: relations between agents and semantic objects (i.e. )

(cf. Partee 1973; Recanati 2000; Potts 2004)

- In a direct report, the quote is an indecomposable unit – a single ex- pression that denotes a linguistic expression (or in certain cases, a non- linguistic representation). It is “mentioned” rather than used.

- In an indirect report, the syntactic/semantic parts of the quote contribute to the syntactic structure and meaning of the whole sentence in the same way as they would in a matrix (non-quote) environment.

Stanford University Page 10 Perspectives in Reported Discourse David Y. Oshima

- An indirect report is syntactically/semantically integrated to the embed- ding clause, in the sense that there is no heterogeneity between the em- bedding and embedded clauses in terms of their syntactic and semantic statuses.

(2) Thesis of Integrity: An indirect quote is integrated to the em- bedding environment, while a direct quote is not.

Stanford University Page 11 Perspectives in Reported Discourse David Y. Oshima

- It has been discussed that certain languages have a mode (or modes) between the direct and indirect modes (semi-indirect mode), which has some characteristics of the direct mode and some of the indirect mode (Coulmas 1985b; Kuno 1988; Kamada 2000; Rice 1985; Reesink 1993; Evans 2005).

- That the binary distinction is not fine-grained enough does not neces- sarily mean that the direct/indirect distinction is continuous (not categor- ical), nor that the thesis of integrity must be compromised.

Stanford University Page 12 Perspectives in Reported Discourse David Y. Oshima

direct… semi-indirect … indirect NOT: continuous scale

direct (not integrated) indirect (integrated) BUT: more “direct-like” indirect reports

more “indirect-like” more “indirect-like” direct reports (?) indirect reports

Stanford University Page 13 Perspectives in Reported Discourse David Y. Oshima

“Direct discourse-like” features in indirect discourse

- Indirect attitude reports can be subdivided along several dimensions, depending on how much they reflect the secondary agent’s perspective.

- I take up four types of such perspectival dimensions:

(3) (i) analytic perspective (de dicto vs. de re attitudes) (ii) logophoric perspective (de se vs. non-de se attitudes) (iii)deictic perspetive (the primary vs. secondary deictic refer- ence point) (iv)empathic perspective (the primary vs. secondary empathy hi- erarchy)

Stanford University Page 14 Perspectives in Reported Discourse David Y. Oshima

De re vs. de dicto attitudes

- Generally, indirect quotes (or descriptive terms within them) have two kinds of interpretations: de re and de dicto.

(4) a. the de dicto interpretation: the quote (or its translation) has the form which the reported agent used or would use to ex- press the quoted utterance/thought, etc. (except for deictic expressions). b. the de re interpretation: the form of the quote can be altered in a way that the quoted speaker would not necessarily ac- cept.

Stanford University Page 15 Perspectives in Reported Discourse David Y. Oshima

(5) Oedipus said that his mother was beautiful.

a. Oedipus said: “My mother is beautiful”. (de dicto) b. Oedipus said: “My wife/Jokasta is beautiful”. (de re)

- The de re/de dicto distinction can be construed as a matter of the speaker’s perspective (Coulmas 1985b). In a de re report, the exter- nal speaker takes the liberty of introducing additional information, which is not available to the quoted agent, into the quote. In a de dicto report, on the other hand, the external speaker sticks to “choose the words” from the quoted agent’s perspective.

- I will use the term analytic perspective to refer to the point of view in this sense, relative to which “choice of words” is made.

Stanford University Page 16 Perspectives in Reported Discourse David Y. Oshima

De se vs. non-de se attitudes

- When a reported utterance/belief involves to the secondary agent (the reported agent) or other coordinates of the secondary (time, place, etc.), the report has either the de se or non-de se interpre- tation, or is ambiguous between the two.

Stanford University Page 17 Perspectives in Reported Discourse David Y. Oshima

- The following report is ambiguous (Percus and Sauerland 2003a):

(6) Pavarottii believes that hisi pants are on fire. a. Pavarotti is disposed to say: “My pants are on fire.” (de se) b. Pavarotti may not be disposed to say: “My pants are on fire.” (non-de se)

In a situation where Pavarotti is looking at his image in a mirror without being aware that the man in the mirror is himself, and sees the man’s pants be on fire (and says to himself: “Gee, that guy’s pants are on fire”), (6) is false on the de se interpretation while true on the non-de se inter- pretation.

Stanford University Page 18 Perspectives in Reported Discourse David Y. Oshima

- Many languages have anaphoric expressions called secondary indexi- cals (or quasi-indexicals), which single out the de se interpretation. (Schlenker 2003; Oshima in press):

(i) Logophoric pronouns (Ewe, etc.)

a. Kofii be ye`i/∗j-dzo. Kofi say log-leave ‘Kofii said that hei left.’

b. Kofii be e∗i/j-dzo. Kofi say 3sg-leave ‘Kofii said that hej/shej left.’

Stanford University Page 19 Perspectives in Reported Discourse David Y. Oshima

(ii) Long distance reflexives (Icelandic, etc.)

a. Jon´ i syndi´ Haraldij fot¨ a´ sigi/j.

‘Johni showed Haroldj clothes for himselfi/j.’

b. Jon´ i segir aδ Mar´ıa elski sigi/∗j.

‘Johni says that Mary loves (subj.) himi.’

(iii) “All-purpose” indexicals (Amharic, etc.)

Johni says that Ii am a hero.

‘Johni says that hei is a hero.’

(iv) Ambiguous pronouns? (English, etc.)

Johni says that {he*i/hei} is a hero.

Stanford University Page 20 Perspectives in Reported Discourse David Y. Oshima

- In certain cases, reference to the secondary agent can be made either with a secondary indexical or a regular pronoun:

(7) a. Johni-wa [zibuni-ga boku-o tasuke-ta] to omotte-i-ru.

‘Johni believes that selfi helped me.’ (de se)

b. Johni-wa [karei-ga boku-o tasuke-ta] to omotte-i-ru.

‘Johni believes that hei helped me.’ (non-de se)

(8) a. Situation 1: John believes: “I helped David” b. Situation 2: Amnesic John believes: “The guy called John helped David”

(7a) is compatible only with Situation 1; (7b) is compatible with both Sit- uation 1 and 2.

Stanford University Page 21 Perspectives in Reported Discourse David Y. Oshima

- In certain other cases, as a rule reference to the secondary agent must be made with a secondary indexical:

(9) a. Johni believes that he*i helped me. (he* = logophoric pronoun)

b. Johni believes that hei helped me.

(10) a. Situation 1: John believes: “I helped David” b. Situation 2: Amnesic John believes: “The guy called John helped David”.

In Bafut, (9b) is possible only in an unusual situation like (10b), and is not compatible with (10a).

Stanford University Page 22 Perspectives in Reported Discourse David Y. Oshima

- The occurrence pattern of secondary indexicals (under ‘believe’) in Ba- fut parallels that of primary indexicals.

(11) a. I (referring to the speaker) am smart. b. #He (referring to the speaker) is smart.

- The pattern in Japanese suggests that in certain cases the speaker can choose whether or not to introduce the secondary context relative to which (secondary or all-purpose) indexicals are interpreted.

- The presence of such a secondary context indicates that the speaker describes the quoted material referring to the reported agent’s perspec- tive; I will call this the logophoric perspective.

Stanford University Page 23 Perspectives in Reported Discourse David Y. Oshima

- Which of the primary or secondary logophoric perspective is allowed is to some extent lexically determined (Schlenker 2003; Anand and Nevins 2004):

(12) a. Certain attitude predicates require the secondary logophoric perspective, i.e., require that an indexical be used to refer to the secondary agent (e.g. Bafut ‘believe’, Slave ‘tell’), b. Some others require the primary logophoric perspective, i.e., exclude secondary indexicals from the complement clause (e.g. Slave ‘know’, Zazaki attitude verbs except ‘say’) c. The others allow the alternation between indexicals and or- dinary third person pronouns (e.g. Japanese attitude predi- cates in general, Zazaki ‘say’).

Stanford University Page 24 Perspectives in Reported Discourse David Y. Oshima

Deictic reference point

- In a matrix environment, the acceptability of deictic motion verbs go and come is conditioned by the location of the speaker/addressee (Fillmore 1997).

(13) (Situation: The speaker is not at John’s office at the utterance time, and the addressee is not at John’s office at the utterance time or at the event time.) I {went/#came} to John’s office.

Stanford University Page 25 Perspectives in Reported Discourse David Y. Oshima

- The conditions under which go and come are used can be formulated as follows (among other ways):

(i) go and come refer to a set of contextually salient individuals: RP (reference points).

(ii) a. go requires that no member of RP be at the goal of motion at the utterance time. b. come requires that (i) there be some member of RP at the goal of motion at the utterance time or at the event time, or (ii) the goal be the home base of a member of RP.

Stanford University Page 26 Perspectives in Reported Discourse David Y. Oshima

(iii) The selection of the members of RP is conditioned by the follow- ing rules (in English):

a. The speaker is always a member of RP. b. It is preferred for the addressee to be a member of RP as well (e.g. {?John/?I} went to see you). The degree of preference is affected by various factors; under certain conditions, the inclusion of the addressee to RP is almost obligatorily (e.g. ??Should I go to see you?) c. When neither the speaker nor the addressee is the theme (the moving entity), a third person individual can be chosen as a member of RP (e.g. John came to visit Mary).

Stanford University Page 27 Perspectives in Reported Discourse David Y. Oshima

- In an embedded environment, the choice of reference point(s) does not follow the rules summarized above:

(14) John believes that I {went/came} to his office. (cf. John was writing a letter when I {went/#came} to his office.)

- The between (13) and (14) suggests that in an embedded en- vironment the reference point for deictic predicates can be chosen either from the external speaker’s or the secondary agent’s deictic perspec- tive.

Stanford University Page 28 Perspectives in Reported Discourse David Y. Oshima

Empathy hierarchy

- Certain languages have altering forms of transitive/ditranstive verbs:

(15) a. Direct forms: the speaker’s perspective is closer to the sub- ject (the speaker empathizes more with the subject than with the object). b. Inverse forms: the speaker’s perspective is closer to the (di- rect or indirect) object (the speaker empathizes more with the object than with the subject).

(Oshima to appear, Kuno 1987)

Stanford University Page 29 Perspectives in Reported Discourse David Y. Oshima

- Japanese giving verbs yaru and kureru form such a pair (yaru = direct, kureru = inverse; note that they share the same argument structure):

(16) a. Boku-wa Taro-ni prezento-o {yaru/*kureru}. I-Top Taro-Dat gift-Acc give-Pres ‘I will give Taro a gift.’ b. Taro-wa boku-ni prezento-o {*yaru/kureru}. Taro-Top I-Dat gift-Acc give-Pres ‘Taro will give me a gift.’

- In a matrix environment, it is impossible to choose a third person par- ticipant as the empathy locus when the other argument is the speaker.

Stanford University Page 30 Perspectives in Reported Discourse David Y. Oshima

- In an embedded environment, it appears that this constraint can be violated:

(17) a. Taro-wa [boku-ga kare-ni prezento-o yatta/kureta] to Taro-Top [I-Nom he-Dat gift-Acc give-Past] Quot omotte-i-ru. believe-Asp-Pres. ‘Taro believes that [I gave him a gift].’ b. Taro-wa [kare-ga boku-ni prezento-o yatta/kureta] to Taro-Top [he-Nom I-Dat gift-Acc give-Past] Quot omotte-i-ru. believe-Asp-Pres. ‘Taro believes that [he gave me a gift].’

Stanford University Page 31 Perspectives in Reported Discourse David Y. Oshima

- This suggests that in an embedded environment the empathy hierarchy (the ranking of individuals based on the degree to which the speaker em- pathizes with them) can be determined either from the external speaker’s or the secondary agent’s empathic perspective.

Stanford University Page 32 Perspectives in Reported Discourse David Y. Oshima

Semantic analyses of the four perspectival dimensions (outline)

(18) (i) analytic perspective (de dicto vs. de re attitudes) (ii) logophoric perspective (de se vs. non-de se attitudes) (iii)deictic perspetive (the primary vs. secondary deictic refer- ence point) (iv)empathic perspective (the primary vs. secondary empathy hi- erarchy)

Stanford University Page 33 Perspectives in Reported Discourse David Y. Oshima

Analytic Perspective

- Previous approaches:

• The classical approach De dicto attitudes = descriptive terms taking narrow scope; De re attitudes = de- scriptive terms taking wide scope

• Kaplan’s approach based on the “acquaintance” relation A term in an can be related to (“quantify into”) a reported ut- terance/belief only if there is a name (descriptive term) that the subject (secondary agent) associates with it.

Stanford University Page 34 Perspectives in Reported Discourse David Y. Oshima

- A model of attitude reports along the lines of Crimmins and Perry (1989):

(i) An attitude report is about a relation between an agent and a cognitive entity (“a thing in the agent’s head”) that has a propo- sition as its content (a mental representation of state of affairs; MRSOA). An MRSOA “represents” a proposition.

(ii) An agent can have various attitudes toward his MRSOAs.

(iii) MRSOAs (thoughts) are structured entities that contain notions and ideas as constituents. Notions are cognitive entities that are (causally) related to individuals; ideas are, similarly, cogni- tive entities related to properties/relations (or more higher-order objects). Notions and ideas represent individuals and proper- ties/relations, respectively.

Stanford University Page 35 Perspectives in Reported Discourse David Y. Oshima

m(rsoa)1, m2, … i(dea) , i , … 1 2 P = R(a, b) causal connection n(otion)1, n2, … (“content of” relation) Structure (m1) =

Content (m1) = P

m1

attitude relation (believe, say, fear, desire, …)

Stanford University Page 36 Perspectives in Reported Discourse David Y. Oshima

(19) Lois Lane believes that Superman is strong.

(20) λw1[∃m1[believe(lois.lane, m1, w1) ∧

Content(m1) = λw2[strong(superman, w2)]]]

(n.b.) Truth Definition: An expression λw1[α] that serves as a translation of a nat- ural language matrix sentence is true in the context c in the world w under the assignment g iff [[α ]]c,w,g[w1→w] = 1.

Stanford University Page 37 Perspectives in Reported Discourse David Y. Oshima

- An alternative analysis of de dicto attitudes: for ‘A v’s S’ to be true de dicto, A must be disposed to express the content of the relevant MRSOA by some sentence S’, which is identical to S except that all deictic terms in S are replaced by certain other names.

(21) Lois Lane believes that Clark Kent is strong. (de dicto)

(22) λw1[∃m1[∃u1[believe(lois, m1, w1) ∧

Content(m1) = λw2[strong(clark.kent, w2)] ∧ d e assent(lois, m1, u1, w1) ∧ u1 ∈ DA( Clark Kent is strong )]]]

(n.b.) (i) The function DA (“deictic adjustment”), applied to a sentence S, yields a set of sentences identical to S except that all occurrences of indexicals/pronominals (I, you, this, ...) are replaced by certain other terms. (ii) [[assent(a, m, u, w)]]w,c,g = 1 iff [[a]]w,c,g assents that [[m]]w,c,g can be linguistically expressed by [[u]]w,c,g in [[w]]w,c,g.

Stanford University Page 38 Perspectives in Reported Discourse David Y. Oshima

Logophoric Perspective

- It has been widely acknowledged that the object of a de se attitude must be an entity that is more fine-grained than a proposition (propositional content, a set of possible worlds), such as:

• property (Lewis 1979; Chierchia 1989)

• a set of centered possible worlds (individual-world pairs) (Cresswell and von Stechow 1982)

• propositional character (a function from contexts, i.e., tuples of a speaker, an addressee, a time, a place, ... , to propositional contents) (Schlenker 2003)

Stanford University Page 39 Perspectives in Reported Discourse David Y. Oshima

- The same idea can be easily incorporated to the MRSOA-based analy- sis of attitude reports.

- Some MRSOAs (thoughts) are indexical in their nature – Pavarotti would behave differently when he has an ‘I’ thought (‘My pants are on fire’) and when he has a ‘he’ thought (‘His pants are on fire’).

- What an MRSOA represents is a propositional character, rather than a propositional content.

Stanford University Page 40 Perspectives in Reported Discourse David Y. Oshima

(23) I am rich 7→ λw1[rich(Ag(c∗), w1)]

(24) Johni-wa [zibuni-ga kanemoti-da] to omotte-i-ru. John-Top self-Nom rich-be.Pres Quot believe-Asp-Pres ‘Johni believes that he*i is rich.’

(25) (24) 7→ λw1[∃m1[believe(john, m1, w1) ∧ Character(m1) =

λc1[λw2[rich(Ag(c1), w2)]]]]

(n.b.) a. [[c∗]]w,c,g is defined only if g(c∗) = c. If defined, [[c∗]]w,c,g = g(c∗). w,c,g b. [[ci]] = g(ci).

Stanford University Page 41 Perspectives in Reported Discourse David Y. Oshima

- The complement-clause of an attitude report can be content-denoting, rather than character-denoting.

(26) Johni-wa [karei-ga kanemoti-da] to omotte-i-ru. John-Top he-Nom rich.be-Pres Quot believe-Asp-Pres ‘John believes that he is rich.’

a. John believes: ‘I am rich’. b. Amnesic John, after reading his own biography, believes: ‘This guy called John is rich.’

The difference between the two characters ((26a) and (26b)) is “neutral- ized” in the report.

Stanford University Page 42 Perspectives in Reported Discourse David Y. Oshima

- When the complement clause is character-denoting (n.b. the event ar- gument is added to each ) ...

(24) 7→ λw1[∃e1[∃m1[believe(john, m1, e1, w1) ∧ Character(m1) =

λc1[λw2[∃e2[rich(Ag(c1), e2, w2)]]]]]]

- When the complement clause is content-denoting ...

(26) 7→ λw1[∃e1[∃m1[believe(john, m1, e1, w1) ∧ Character(m1) ∈ 0 0 {χhc,hs,tii: χ (hjohn, Timee(e1), Placee(e1)i) = λw2[∃e2[rich(john, e2, w2)]]}]]]]

Stanford University Page 43 Perspectives in Reported Discourse David Y. Oshima

Deictic Perspective

- In a semantic model, the set of reference points (RP) for go/come can be defined as an index relative to which an expression is evaluated (on a par with the world, context, and variable assignment).

Stanford University Page 44 Perspectives in Reported Discourse David Y. Oshima

(27) go 7→ λpl1[λx[λw1[move-to(x, pl1,w1){¬∃y[y ∈ r1 ∧ at(y,

pl1, w1)]}]]]]

(28) come 7→ λpl1[λx[λw1[move-to(x, pl1, w1){∃y[y ∈ r1 ∧ at(y,

pl1, w1)]}]]] where,

(29) [[φ{ψ}]]w,c,r,g is defined only if [[ψ]]w,c,r,g = 1. If defined, [[φ{ψ}]]w,c,r,g = [[φ]]w,c,r,g.

- A free RP variable ri always refers to the RP of the context of utterance:

w,c,r,g (30) An RP variable ri must be bound unless [[ri]] = r

(i.e. g is such that g(ri) = r)

Stanford University Page 45 Perspectives in Reported Discourse David Y. Oshima

(31) a. John believes that I went to his office. (the deictic perspective anchored to the speaker) 7→

λw1[∃m1[believe(john, m1, w1) ∧ Content(m1) =

λw2[move-to(Ag(c∗), j’s-office, w2){¬∃x[x ∈ r1 ∧ at(x, j’s-office, w2)]}]]] b. John believes that I came to his office. (the deictic perspec- tive anchored to John) 7→

λw1[∃m1[∃r1[believe(john, m1, w1) ∧ adopt dx(john, r1, w1) ∧ Content(m1) =

λw2[move-to(Ag(c∗), j’s-office, w2){∃x[x ∈ r1 ∧ at(x, j’s-office, w2)]}]]]] where,

(32) [[adopt dx(a, r, w)]]w,c,r,g = 1 iff [[r]]w,c,r,g is the set of refer- ence points for [[a]]w,c,r,g in [[w]]w,c,r,g.

Stanford University Page 46 Perspectives in Reported Discourse David Y. Oshima

Empathic Perspective

- Empathy-loaded expressions are interpreted with respect to the contex- tually adopted empathy hierarchy, parallel to deictic motion verbs expres- sions that are interpreted with respect to the contextually determined RP.

- The empathy hierarchy is a partially ordered set of individuals; the or- dering reflects the relative degree to which the speaker empathizes with its members.

- That the speaker adopts a hierarchy h within which a outranks b is equivalent to that the speaker empathizes with a more than with b.

Stanford University Page 47 Perspectives in Reported Discourse David Y. Oshima

(33) a. yaru 7→ λz[λy[λx[give(x, y, z){¬outrank(z, x, h1)}]]]

b. kureru 7→ λz[λy[λx[give(x, y, z){outrank(z, x, h1)}]]] where,

(34) [[outrank(a, b, h)]]w,c,r,h,g = 1 iff [[a]]w,c,r,h,g outranks [[b]]w,c,r,h,g within [[h]]w,c,r,h,g

- Again, a free EH variable hi always refers to the empathy hierarchy adopted by the external speaker.

w,c,r,h,g (35) An EH variable hi must be bound unless [[hi]] = h

(i.e. g is such that g(hi) = h)

Stanford University Page 48 Perspectives in Reported Discourse David Y. Oshima

(36) Johni-wa [boku-ga karei-ni hon-o {a. yat/b. kure}-ta] to John-Top I-Nom he-Dat book-Acc give-Past Quot omotte-i-ru. believe-Asp-Pres ‘John believes that I gave him the book.’ (the empathic perspective anchored to {a. the speaker/b. John})

(i) (36a) 7→ λw1[∃m1[believe(john, m1, w1) ∧ Content(m1) = λw2[give(Ag(c∗),the-

book,john){¬outrank(john, Ag(c∗), h1)}]]]

(ii) (36b) 7→ λw1[∃m1[∃h1[believe(john, m1, w1) ∧ adopt emp(taro, h1, w1)

Content(m1) = λw2[give(Ag(c∗),the-book,john){¬outrank(john, Ag(c∗),

h1)}]]] where,

(37) [[adopt emp(a, h, w)]]w,c,r,h,g = 1 iff the empathy hierarchy for [[a]]w,c,r,h,g in [[w]]w,c,r,h,g is [[h]]w,c,r,h,g

Stanford University Page 49 Perspectives in Reported Discourse David Y. Oshima

Discussion: What determines the perspective?

- What determines the speaker’s choice of each type of perspective? Are there factors that interact with all types of perspectives?

- Do the types of perspectives have to, or at least tend to, match? Is “split perspective” possible?

Stanford University Page 50 Perspectives in Reported Discourse David Y. Oshima

Vividness

- I suggested earlier that an indirect report where one or more perspec- tival feature is shifted (anchored to the secondary agent) is more “direct- discourse like”.

- Certain communicative factors that motivate the speaker to choose the direct mode, rather than the indirect mode, might be relevant to the speaker’s choice of the perspective in indirect discourse as well.

- In particular: It has been said that direct reports are more “vivid”, “lively”, and “theatrical” than indirect reports (Tannen 1985; Clark and Gerrig 1990). The choice of the secondary (shifted) perspective may indicate the speaker’s intention to make the report vivid, lively, and theatrical.

Stanford University Page 51 Perspectives in Reported Discourse David Y. Oshima

Predicate types

- As we have seen, some attitude predicates require/prefer the primary lo- gophoric perspective, while others require/prefer the secondary logophoric perspective.

- The correlation between semantic types of predicates and the primary/secondary logophoric perspective is captured by what is known as the universal im- plicational hierarchy of logophoric predicates (Culy 1994; Stirling 1993):

(38) Speech Predicates (say) > Epistemic/Psychological Predicates (believe, fear) > Knowledge Predicates (know) > Perceptive Precicates (see)

Verbs belonging to a higher class in the hierarchy tend to favor the sec- ondary logophoric perspective.

Stanford University Page 52 Perspectives in Reported Discourse David Y. Oshima

- The same hierarchy seems to have similar effects on other perspectival dimensions:

• deictic

(39) a. John said to Mary that he needed to {go/?come} to my place. b. John believed that he needed to {?go/come} to my place. c. John knew that he needed to {??go/come} to my place.

Stanford University Page 53 Perspectives in Reported Discourse David Y. Oshima

• empathic

(40) Boku-wa ano gakusei-ni suisenzyoo-o kaite-yar/*kure-(a)nakat-ta. ‘I did not write a letter of recommendation for that student.’

a. Ano gakusei-wa [boku-ga suisenzyoo-o kaite-?yar/kure-(a)nakat-ta] to minna- ni iihurasite-i-ru. ‘That student is telling everyone that I did not write a letter of recommen- dation for him.’ b. Ano gakusei-wa [boku-ga suisenzyoo-o kaite-yar/kure-(a)nakat-ta] to sinzikonde- i-ru. ‘That student (wrongly) believes that I did not write a letter of recommen- dation for him.’ c. Ano gakusei-wa [boku-ga suisenzyoo-o kaite-yar/*kure-(a)nakat-ta] koto- o okkotte-i-ru. ‘That student is angry that I did not write a letter of recommendation for him.’

Stanford University Page 54 Perspectives in Reported Discourse David Y. Oshima

• analytic (?)

(41) a. John said that Samuel Clemens was a great author (but he did not say that Mark Twain was). b. John believes that Samuel Clemens is a great author (but he does not believe that Mark Twain is). c. John knows that Samuel Clemens is a great author (?*but he does not know that Mark Twain is).

Stanford University Page 55 Perspectives in Reported Discourse David Y. Oshima

- Possible explanations:

(42) a. Belief “reports” necessarily involve the (re)construction of the reported material by the external speaker, and this makes the external speaker’s perspective predominant. b. Knowledge/perception reports require that the external speaker conceive the reported content as a fact (have first-hand ac- cess to the reported content), and thus favor the primary per- spective.

Stanford University Page 56 Perspectives in Reported Discourse David Y. Oshima

Interaction among perspectives

- Given an indirect report which involves two (i.e. primary and secondary) agents, and where the attitude predicate can be either character-taking or content-taking, theoretically there are 24 = 16 possible patterns of perspective-taking:

(43) Analytic Logophoric Empathic Deictic 1. Primary Primary Primary Primary 2. Primary Primary Primary Secondary . . 16. Secondary Secondary Secondary Secondary

Stanford University Page 57 Perspectives in Reported Discourse David Y. Oshima

- Are the four dimensions of perspectives totally independent from each other?

(44) a. “Independence” Hypothesis: A perspectival dimension can be shifted or remain unshifted regardless of whether other dimensions are shifted or not. b. “Co-variation” Hypothesis: A perspectival dimension is shifted iff the other dimensions are shifted, and remains unshifted iff the other dimensions are unshifted. c. “Implicational Hierarchy” Hypothesis: A perspectival dimension can be shifted only if other dimensions which are lower than it in the implicational hierarchy are shifted.

- To determine the relations between the four dimensions, it will be crucial to look into data from a language where shift of all the dimensions is possible.

Stanford University Page 58 Perspectives in Reported Discourse David Y. Oshima

(i) The analytic dimension is largely independent from the other three dimensions; the correlation is weak, if not absent.

(45) Taroi-wa [boku-ga karei-no tokoro-ni Spiderman-no kantoku-o Taro-Top I-Nom he-Gen place-Dat Spiderman-Gen director-Acc turete-{a. iku/b. kuru} -beki-da] to omotte-i-ru. take-{go/come} -should Quot believe-Asp-Pres (a) ‘Taroi thinks that I should go to where hei is with the director of Spiderman.’

(b) ‘Taroi thinks that I should come to where hei is with the director of Spider- man.’

Stanford University Page 59 Perspectives in Reported Discourse David Y. Oshima

(ii) An implicational relation holds between the logophoric, deictic, and empathic perspective:

(46) empathic perspective > logophoric perspective > deictic per- spective

(a) Logophoric Deictic A1 Primary Secondary acceptable A2 Secondary Primary marginal

(b) Empathic Logophoric B1 Primary Secondary more or less acceptable B2 Secondary Primary marginal

(n.b.) primary = unshifted, secondary = shifted

Stanford University Page 60 Perspectives in Reported Discourse David Y. Oshima

(iii) The mismatch between the deictic and empathic perspective leads to unacceptability, no matter what the “direction” of split is.

(c) Empathic Deictic C1 Primary Secondary at best marginal C2 Secondary Primary at best marginal

Stanford University Page 61 Perspectives in Reported Discourse David Y. Oshima

(A1) Taroi-wa [boku-ga karei-no tokoro-ni kuru-beki-da] to syutyoo-site-i-ru. Taro-Top I-Nom hei-Gen place-Dat come-should Quot claim-Asp-Pres ‘Taroi claims that I should come to where he*i is.’

(A2) ??Taroi-wa [boku-ga zibuni-no tokoro-ni iku-beki-da] to syutyoo-site-i-ru. Taro-Top I-Nom selfi-Gen place-Dat go-should Quot claim-Asp-Pres ‘Taroi claims that I should go to where he*i is.’

(B1) (?)Taroi-wa [boku-ga zibuni-ni sono hon-o yat-ta] to syutyoo-site-i-ru. Taro-Top I-Nom selfi-Dat that book-Acc give-Past Quot claim-Asp-Pres ‘Taroi claims that I gave him*i that book.’

(B2) ??Taroi-wa [boku-ga karei-ni sono hon-o kure-ta] to syutyoo-site-i-ru. Taro-Top I-Nom hei-Dat that book-Acc give-Past Quot claim-Asp-Pres ‘Taroi claims that I gave himi that book.’

(C1) ?*Taroi-wa [itumo boku-ga φi yuubinbutu-o todoke-ni kite-yatte-i-ru] to syutyoosite-i-ru. Taro-Top always I-Nom φ mail-Acc hand-Dat come-Ben-Asp-Pres Quot claim-Asp-Pres ‘Taroi claims that I always come to deliver mail to himi.’

(C2) ?*Taroi-wa [itumo boku-ga φi yuubinbutu-o todoke-ni itte-kure-i-ru] to syutyoosite-i-ru. Taro-Top always I-Nom φ mail-Acc hand-Dat go-Ben-Asp-Pres Quot claim-Asp-Pres ‘Taroi claims that I always go to deliver mail to himi.’

Stanford University Page 62 Perspectives in Reported Discourse David Y. Oshima Summary

• Some indirect reports are more “direct discourse-like” in the sense that certain perspectival features are anchored to the secondary agent (reported speaker).

• The mental representation-based model of attitude reports (a` la Crim- mins and Perry) has certain advantages over the Hintikkian doxastic alternatives-based model; among other things, it provides a basis for the “sententialist” analysis of de dicto attitudes.

• Whether a perspectival feature is anchored to the primary or sec- ondary agent is affected/conditioned by: (i) the semantic class of the attitude predicate, and (ii) to whom other perspectival features are an- chored.

Stanford University Page 63