Freitag, 24. Oktober 2008

la meva nova frase favorita...

... l'he treta del Journal intime del poeta i filòsof suís Henri-Frédéric Amiel:

«
Qui veut voir parfaitement clair avant de se déterminer ne se détermine jamais. Qui n’accepte pas le regret n’accepte pas la vie.»

fins el mai


juraria que havia tingut un amic o amiga a nyc, però ara resulta que somiava. hauria jurat que era real, que havia tingut converses amb aquest ens la cara del qual es desdibuixa com les altres formes del somni, juraria que ens havíem petat de riure, i que ens havíem plorat mutuament, i que ens havíem salvat mutuament en moltes ocasions, i que en certa manera aquesta vida acadèmica es feia menys sinistra sabent que hi havia aquest ser misteriós a qui podies parlar... però bé, de vegades sembla que no podem distingir els somnis de la realitat, les ombres de la caverna de la realitat del món de les idees... la imaginació de la ficció... és trist veure com el silenci omple sense despentinar-se el que havia semblat una conversa animada, i és més trist adonar-se que tot havia estat un somni, o potser eres tu qui era al somni d'aquella altra persona que ara ja s'ha despertat i ja no se'n recorda, i ets tu el que s'esborra i se'n va per l'aiguera, i ets tu qui ja no existeix. aquesta no és una carta d'amor, és una carta a aquell amic o amiga del somni, que s'enfonsa o m'enfonsa per sempre més en les capes de glia més remotes del laberint neuroanal, on mai arriba la llum. em pregunto per què collons hem de somiar coses que no existeixen? és realment molt poc pràctic... proposaré que es faci una llei per regular aquests abusos

back in exile


Està confirmat: el meu advisor Chris Barker rebutja continuar treballant amb mi en el segon QP. Els seus motius són, bàsicament, que li sembla un tema massa gran, i massa fora dels meus actuals coneixements com per ser un tema que es pugui encabir en un QP de tres mesos. La seva inflexibilitat ha estat manifesta, i no han servit de res les meves protestes contra l'afusellament per un crim que encara no havia comès. En definitiva, que ja torno a no tenir advisor. Veurem què passa en els propers dies. Mentrestant, us faig un resum del que volia fer, perquè podeu jutjar si el que jo déia estava tan fora de lloc com aparantment ell suposa. I gràcies pel suport dels vikings i les meladoris.

LA PROPOSTA

(i) In this QP, I want to provide a semantic analysis of the linguistic meaning of DOM (a-marking) in Spanish. (ii) I want to argue against the idea that DOM marks specificity in Spanish (LEONETTI 1999, 2006, BLEAM 1999, BRUGÈ & BRUGGER 1996, LACA 1987, TORREGO 1998, 1999). And (iii) I want to argue in favor of the idea that DOM individuates the DP it selects (via case-marking with accusative), and makes it referentially fix (i.e. DOM marks wide scope), and that this is independent of specificity, which I define as epistemic specificity, in line with proposals in ENÇ 1991, FARKAS 1995, KAGAN 2006, and KAMP & BENDE-FARKAS 2006.

(1) Dos leones (se) comieron una cebra (partitive or caseless: narrow scope)
y un tigre dos.
(2) Dos leones (*se) comieron a una cebra (accusative case: wide scope)
y un tigre *(a) dos

(iv) THE BIG PICTURE: I propose a new treatment of specificity as a property of referential DPs only, which is undefined for quantificational DPs, and in consequence, I reject the separation between scopal and epistemic specificity in FARKAS 1995. There’s no such thing as scopal specificity: scopal specificity is scope, period.

A SIDE (NON-UNRELATED) COMMENT

-Newton had an imperfect idea about what gravity was. In an appendix to his Philosophia Naturalis Principia Mathematica, Newton wrote his famous sentence: hypotheses non fingo (I feign no hypotheses) as an answer to those who had publicly challenged him to give an ex-planation for the causes of gravity rather than just the mathematical principles of kinetics.

Along the same lines, biologist work today assuming an idea of organism, although there is not a unified concept of what an organism is.

Arithmeticians work in the properties of numbers even though there is no one encompassing definition of number and the concept of number is open for further development.

Or we can indeed talk about the concept of language, or meaning, and many others within linguistics, which are far from being clearly defined or unified in the literature.

What all this means is that many times scientists have to assume a definition of a fundamental concept, even though it is not perfect, and work from there.

This should apply to this paper for specificity. I propose differences in truth-conditions (see below), and I assume a definition of the concept that I haven’t made up myself, but rather adopted from the literature (FARKAS, VON HEUSINGER, LEONETTI, and so on). I think I should build my paper from that assumption instead of being stuck because I don’t have a perfect (or perfectly formalized) definition of specificity. As the Swiss poet and philosopher Henri-Frédéric Amiel says in his Journal intime, “the man who insists upon seeing with perfect clearness before he decides, never decides”. Sometimes we have to move on, even in the absence of a perfect foundational basis, because science split from philosophy in the 17th century (starting with Galileo and Newton) on the basis of that idea, and we linguists should be scientists, rather than philosophers.


TYPES OF INDEFINITES ACCORDING TO MY THEORY:

- Specific: there is a fix reference known by some element of the sentence (absolute specificity, known by the speaker), relative specificity (known by some other element of the sentence).
- Non-specific: there is a fix reference, but it is not known by the elements of the sentence, including the speaker as part of those elements.
- Quantificational: there is no fix reference, because the indefinite is only about a quantity.


How do we define reference? As the denotation of a term. How do we define specificity? In terms of the cognitive state of some individual that appears represented in the sentence, including the speaker.

Consider the following interpretations of the indefinite two students in (1):

(1) Two students have to pick up three speakers (each) at the airport
a. There are two students (a, and b) such that each one has to pick up three speakers.
b. Two students (we don’t know which ones) have to pick up three speakers each.
c. There are three speakers, such that each one will be picked up by two students.

(a) corresponds to the standard wide scope reading of two students (scopal specificity, in the sense of FARKAS 1995): specific referential reading. Fix and known reference that makes the interpretation of the other indefinite co-vary: there are six speakers, and two students.
(b) corresponds to a wide scope reading of two students, but with unknown reference, i.e. now we don’t know which two students, all we know is that they are two. This is a non-specific referential reading. Fix reference, but not known, that makes the other indefinite co-vary, i.e. we again have two students and six speakers.
(c) corresponds to the standard narrow scope reading of two students, that depends on the inter-pretation of another operator. This is a quantificational reading, and specificity does not apply, because reference is not at issue. There are three speakers, and up to possibly six dif-ferent students.

SOME DEFINITIONS OF SPECIFICITY IN THE LITERATURE:

KLAUS VON HEUSINGER 2002
- Specificity is a referential property of DPs, which cuts across the distinction between definite vs. indefinite (both definite and indefinite DPs exhibit different ways of referring).
- A specific DP indicates that its associated discourse item is referentially anchored to another object in the discourse (its anchor) and therefore inherits the scopal properties of that anchor. Referentially anchored means that the referent of the specific DP is functionally dependent on the referent of another expression. This relation is sentence bound: a specific DP can only be anchored to discourse items that are explicit in the same sentence (or to the speaker of the sentence).
- A DPi in a sentence φ with respect to a file F and the domain of filenames Dom(φ) is [+specific] if there is a contextual salient function ƒ such that i = ƒ(j) and j ∈ Dom(φ).

Donka Farkas 1995
Scopal specificity: the interpretation of scopally specific indefinites is not dependent on any quantifier or intensional predicate.
The value of a scopally specific indefinite (when interacts with intensional predicates) is chosen from the domain of w, the world with respect to which the main clause is evaluated, while the value of scopally non-specific in-definites is to be chosen from the domain of the world or worlds introduced by the predicate. In the latter case, we have non-rigid reference because the value of the indefinite changes from world to world.
Epistemic specif.: the speaker has an intended referent in mind when using a specific indefinite (idea extracted from Fodor & Sag 1982). Epistemically specific indefinites are characterized with respect to the speaker’s epistemic modal base.
Epistemic specificity is characterized in terms of rigidity of reference with respect to a set of worlds, just like scopal specificity, with the difference that the set of worlds in question is present in the context set of the speech act, rather than being introduced by linguistic expressions.
Epistemically specific indefinites do not contribute an existential quantifier, and have a Kaplan-style contextual interpretation. They appear to have wid-est scope, although they do not participate in scopal relations, according to Farkas.

LEONETTI 2006
He uses KvH definition of specific as referentially anchored. No further definition is provided, but still the paper is a good paper and makes good predictions for Spanish.

ENÇ 1991
Specific NPs are partitive and introduce an inclusion relation, which is a weak link to a previ-ously established referent, unlike definite NPs, which establish a strong link (of identity) to a previously established referent. Specific indefinites in Turkish are marked with accusative case, and non-specific indefinites carry no case morphology at all. (Something similar with DOM in Spanish: DOM is an accusative case marker, but accusative marks individuation, rather than epistemic specificity). No formal definition of specificity is provided.

Donnerstag, 16. Oktober 2008

el meu advisor semantic


Jo em pensava que un proposal és un escrit d'intencions que fas de cara a una feina que has de desenvolupar. Fas una proposta, és a dir, una hipòtesi per resoldre un problema. Dius per on podria anar la solució, i quina és la literatura que consideres rellevant per tractar el problema. Després, quan desenvolupes el paper, entres més en detall en el tema, veus quins problemes hi ha, etc. Però pel meu advisor, això no és així. La proposta ja ha de contenir una perfecta formalització del problema i dels conceptes emprats, una perfecta idea del que vols dir, etc. I el que és pitjor, tots els conceptes que utilitzes han d'estar perfectament definits. No importa que la majoria de papers utilitzin els conceptes de manera més loose... res de res. En fi, la veritat és que em deprimeix, Aquí al costat teniu el seu careto. Sembla un tio maco, però darrerament tothom li ha començat a agafar mania perquè es comporta com un malparit. El seu nom és Chris Barker.

Samstag, 11. Oktober 2008

un altre acudit, però sense segones

Va uno de Lepe y le dice al otro: Now we can go back to our system. If we use Farkas' concept of cognitive state (or epistemic modal base) of an individual i, represented by Pi, and understood as the set of propositions that i takes to be true of the actual world w0, that is to say:

(*) Pi = {p ∈ ℘(W): ∃!i ∈ Un [p(i)(w0) = 1]} where ℘(W) is the set of all propositions in M; Un ∈ ℘(U) is the set of participants in a relevant context of utterance; p(i)(w0) is a Skolem function that gives us the set of propositions that i takes to be true in the actual world, w0, i.e. the epistemic modal base of i; and if we take the epistemic state of every individual i as different to that of everyone else in U, which is why I used the ! with the existential quantifier in (*), as an abbreviation for λP∃i [P(i) & ∀j [P(j) → i = j]].

Therefore, ⟦εixFx⟧M, g[i/sp] = Φi (⟦Fx⟧M, g[i/sp]).

Y el otro de Lepe dice: no hijo, no!!!!