Ancient Greek Monsters Corien Bary and Emar Maier University of Nijmegen <{corien|e.maier}@phil.kun.nl> Szklarska Poreba 4, 18th June 2003
looking for monsters (1)
Otto believes it is possible that 3 years ago I was in Amsterdam
The I, in (1) refers to me, the actual speaker of the whole sentence despite the embedding under 3 operators An operator that would shift the reference of I is called a ‘monster’. Do such operators exist?
traditional possible worlds semantics basic Fregean intuition: Sinn = proposition = object of attitudes (like believe) = content = meaning = what is conveyed by an utterance = the argument of modal operators formally: intension = function from possible worlds to extensions • proposition = intension of sentence: kφk : W → {0, 1} • individual concept = intension of singular term: ktk : W → Ind
indexicals • indexical = deictic term = term whose denotation depends on certain features of the context of utterance • demonstratives and pure indexicals
Ancient Greek monsters / 18th June 2003
<{corien|e.maier}@phil.kun.nl>
1
• paradigmatic examples: I, now, yesterday, there, that . . . • . . . but also tense and/or person features carried by inflections on VPs
a problem with indexicals What semantic value is to be associated with e.g. the English indexical I ? (2)
k Ik(w) =? ∈ Dom(W)
Note that English I always refers to the person uttering it, so. . . (3)
k Ik(w) = the person speaking in w
Actually, we need centered worlds (indices) to model situations more fine-grained than possible worlds (4)
I =W ×T ×P i = hwi , ti , pi i
(5)
k Ik(i) = the person speaking at time ti at location pi in wi
Now consider the intension of the NP the speaker : (6)
k the speakerk(i) = the person speaking at time ti at location pi in wi = k Ik(i)
So (7) is logically valid (true in all models at all indices) (7)
I am the speaker
Therefore, (8) is also logically valid (8)
it is necessary that I am the speaker
But it is not necessary that I am speaking
Kaplan’s two obvious principles
Ancient Greek monsters / 18th June 2003
<{corien|e.maier}@phil.kun.nl>
2 Principle 1 the denotation of indexicals depends on the context of utterance, their linguistic meaning determines how Principle 2∗ indexicals are always interpreted as if they had widest scope, i.e. in the actual world of utterance, therefore their denotation cannot be shifted by any linguistic operator
character and content Kaplan introduces 2 levels of meaning, character and content, to
• solve the problem with indexicals • do justice to both principles • C = W × T × P × Ind c = hwc , tc , pc , sc i • k Ik : C → (I → Ind) (character of the term I) • k Ik(c) : I → Ind (content or intension of I as uttered in context c) • k Ik(c)(i) := the speaker of c = sc content = intension = proposition = what is asserted in an utterance character = linguistic meaning
• extensional (0-dimensional) operator, e.g. ¬ ( 1 if kφk(c)(i) = 0, k¬φk(c)(i) = 0 otherwise • modal (1-dimensional) operator, e.g. 2 ( 1 if ∀w0 [kφk(c)(i0 ) = 1], k2φk(c)(i) = 0 otherwise • monster (2-dimensional operator, context-shifter), e.g. ⊕ ( 1 if ∃c0 [kφk(c0 )(ic0 ) = 1], k ⊕ φk(c)(i) = 0 otherwise
Ancient Greek monsters / 18th June 2003
<{corien|e.maier}@phil.kun.nl>
3
prohibition on monsters Kaplan claims that natural languages cannot contain monsters, because monsters would violate principle 2∗ .Consider the formal monster ⊕ (9) (10)
k ⊕ φk(c)(i) = 1 iff ∃c0 [kφk(c0 )(ic0 ) = 1] I am hungry; hungry(I)
Embedding (10) under ⊕ yields the truthconditions in (11) (11)
k ⊕ hungry(I)k(c)(i) = 1 iff iff ∃c0 [k Ik(c0 )(i 0 ) ∈ k hungryk(c0 )(i 0 )] iff c c iff there is a context whose speaker is hungry
The embedded occurrence of I no longer refers to the actual speaker, so allowing operators like ⊕ causes our system to contradict principle 2∗
cognitive significance According to Kaplan (1989) attitudes like belief are not just propositional, but depend on character as mode of representation content = intension = objective content of thought character = cognitive significance = mode of representation of thought
Kaplan’s pants on fire Kaplan, looking at himself on a TV screen, not recognizing himself: (12)
His pants are on fire
After a while he looks down and finds out: (13)
My pants are on fire
Ancient Greek monsters / 18th June 2003
<{corien|e.maier}@phil.kun.nl>
4 He has learned something important by going from (12) to (13), but the difference is purely characterial; the same content is expressed. His attitudes in this case are sensitive to these character differences.
where to look for monsters in natural language?
• attitudes and sentence (type) meanings are individuated by character • plausible candidates might be attitude and speech reports
English speech reports English has two devices to report speech acts
• report in direct discourse, i.e. using quotation marks (14)
David said: “I am hungry”
• indirect discourse, using a says that. . . operator (15)
Davidi said that hei was hungry
The quotation or report in direct discourse appears to be monstrous and seems to violate principle 2∗ . . . But quotation is not a linguistic operator, it marks a use/mention distinction: the words in the quoted sentence are mentioned. (Formally, an operator modeling our English quotation would confuse object- and metalanguage)
how about indirect discourse? Consider an utterance by Pat of (16), directed to me (16)
you are tired
After that utterance took place, I can report it with (17)
Ancient Greek monsters / 18th June 2003
<{corien|e.maier}@phil.kun.nl>
5 (17)
Pat said that I was tired
• the you and present tense are from the direct speech (16) are changed to I and was in the indirect report (17) • some information is lost; the report is also appropriate if a different indexical or proper name that picks out me were used in the original utterance (as long as the same content is expressed, characterial differences don’t matter) • the ocurrences of I and the past tense in was are evaluated in the actual context of utterance of (17) no violations of the 2 principles, no context-shifting monster required, Kaplan (1989:ch. XX) adequate
a case of no retention? (18)
Lègei
å
kat goroc ±c
Íbris t c eÊmi
3.sg.pres sg.nom sg.nom sg.nom 1.sg.pres that insolent I am says the accuser ‘The accuser says that I am insolent’ the sentence in (18) is a report of the discourse in (19) (19)
Íbris t c eÚ
sg.nom 2.sg.pres insolent you are ‘You are insolent’
• the person inflection is altered in the transition from direct to indirect discourse • there is no apparent violation to the principles • the Greek report with construction with that
±c
/
åti
seems to be the exact analogon of the English indirect discourse
a case of full retention (20)
Prìxenoc eÚpen
åti
gajìc
eÊmi
sg.nom 1.sg.pres nom.sg 3.sg.past I am Proxenos said that good ‘Proxenosi said that hei was good’ Ancient Greek monsters / 18th June 2003
<{corien|e.maier}@phil.kun.nl>
6 is a report of an earlier utterance, by Proxenos, of (21) (21)
gajìc
eÊmi
sg.nom 1.sg.pres I am good ‘I am good’
• all linguistic features of the direct discourse are retained • the indexicals (person and tense) do not take wide scope but are evaluated in the context of the reported speech act, requiring a context shift Or, perhaps we should translate (20) as (22), treating this occurrence of marks? (22)
åti
as the equivalent of quotation
Proxenos said: “I am good”
this would make
åti
very ambiguous, between
1.
åti
= that. . . , introducing a report in (non-monstrous) indirect discourse
2.
åti
= :”. . . ”, marking a use/mention distinction
cases of partial retention
(23)
, oÕc êqomen 3.pl.past inf.pres pl.dat pl.dat rel.pl.acc 1.pl.pres they said [to] come with the horses that [we] have ‘Theyi said theyi were coming with the horses theyi had’ eÚpon
(24)
êrqesjai
sÌn
toØc
Ñppoic
, oÕc êqomen 1.pl.pres pl.dat pl.dat rel.pl.acc 1.pl.pres we come with the horses that we have ‘We are coming with the horses we have’ ârqìmeja
• so,
åti
(25)
sÌn
toØc
Ñppoic
not quotation after all, because partial quotation is not an option: *they said they were coming with the horses “we have”
• Kaplan can’t handle such examples, because the occurrence of we in (23) is really monstrous Ancient Greek monsters / 18th June 2003
<{corien|e.maier}@phil.kun.nl>
7
• additional benefit of a monster based account will be that we don’t need to postulate an ambiguity in åti
gr nÜn aÎtoÈ piènai. tän >Alkibidhn âkèleusan s trathgoÈ inf.pres pl.nom 3.pl.past sg.acc sg.acc pl.nom pl.nom [because] now Alkibiades [they] ordered to leave [self] commanders the
(26)
OÉ
s trathgeØn
inf.pres [to] be in command ‘The commandersi ordered Alkibiades to leave saying that th´eyi were now? in command’ The sentence in (26) reports a situation where some new commanders, after telling Alkibiades to leave, utter (27) (27)
aÎtoÈ
gr
nÜn
s trathgoÜmen
pl.nom 1.pl.pres [self] [because] now we are in command ‘(. . . , because) w´e are now in command’
•
in (26) does not refer to the actual speech time, but to the time of the reported speech act (27), it is retained in the transition nÜn
• but this is not a case of quotation because the finite verb of the direct speech is changed into an infinitive in the report • note the monstrous behavior of the now in the English translation. . .
monsters everywhere • Motherese: (28)
a. b.
Not now, mamma’s busy Go ask daddy
• counterpart person shift (29)
a.
Last night I dreamt I was J-Lo and I kissed me
Ancient Greek monsters / 18th June 2003
<{corien|e.maier}@phil.kun.nl>
8 b.
If I were you, I’d hit me
• Schlenker (2003): Amharic attitude ascriptions, logophoric pronouns (Ewe, English), Russian present tense, German logophoric mood (Kon. I),. . . See also (von Stechow 2001) • perhaps quotation or the semi-quotational uses of quotation marks in natural language (Geurts ≤2002) are also best regarded as monstrous operators (giving up Kaplan’s strict analysis in terms of use/mention)
References Geurts, Bart, ≤2002. The pragmatics of quotation. abstract http://www.phil.kun.nl/tfl/bart/ talks/quotabstr.pdf. Kaplan, David. 1989. Demonstratives: an essay on the semantics, logic, metaphysics, and epistemology of demonstratives and other indexicals. In Themes from Kaplan, ed. by Joseph Almog, John Perry, & Howard Wettstein, 481–563. Oxford: Oxford University Press. Schlenker, Philippe. 2003. A plea for monsters. Linguistics and Philosophy 26.29–120. http: //www-rcf.usc.edu/~schlenke/Monsters.doc. von Stechow, Arnim, 2001. Schlenker’s monsters. handout http://vivaldi.sfs.nphil. uni-tuebingen.de/~arnim10/Handouts/Schlenkers.Monster.pdf.
Ancient Greek monsters / 18th June 2003
<{corien|e.maier}@phil.kun.nl>