*!*Ghomeshi, Jila and Diane Massam. 1994. To Appear, Linguistic Analysis
part of a behavior pattern that superficially treats the whole world as a special friend." pg 99
tải về 302.24 Kb.
|
| part of a behavior pattern that superficially treats the whole world as a special friend." pg 99
Discussion of use of this deletion in actual texts here pg 99--. (Interesting, not clear how controlled it is). {Use your phone? (from Dragnet from Jerry Morgan) Jerry morgan: can't answer with ''no you can't" Concludes with an arguemnt for against treating this type of deletion as ``peripheral" or ``pragmatic" *!*Traugott, Elizabeth C. 1995. The Role of the Development of Discourse Markers in A Theory of Grammaticalization. Paper delivered at ICHL XII, University of Manchester, August 1995. Adverbial phrase $>$ sentence adverbial $>$ discourse marker Discusses syntactic adjunct slot. ``In fact all the evidence shows that it is only lexical items in specifgic syntactic constructions that [undergo grammaticalization]" [sent to ET 11/23/95] It occurred to me that your idea of pragmatic strengthening could be extended to account for the observation that most grammaticalization occurs on lexical items in specific syntactic constructions, if you broaden pragmatic strengthening to mean ``context" strengthening: the linguistic context is part of the overall context in which the word appears. Just as non-linguistic implicatures can give rise to conventionalized aspects of meaning over time, so do implicatures arising from the linguistic contexts in which the words occur. *!*Tomasello, Michael and Patricia J. Brooks. 1996 ms (1997 publication?). Early Syntactic Development: A Construction Grammar Approach. M Barrett (Ed) The Development of Language London: UCL Press. Argues that children try to acquire whole linguistic constructions not subjects, objects etc. Four major steps in syntactic development: Holophrases
Children tend to be conservative: Olguin and Tomasello (1993): 25 mo old children were taught 4 novel transitive verbs over a month, each in a different combinatorial configuration: both participants expressed, agent only, patient only or neither argument expressed. Children almost always reproduced the same exact pattern they had heard: they did not know how to mark particpant roles in a general way. They desctibe the acquisition of various constructions: Locatives Datives resultatives ("no experimental studies of resultative construction have been conducted with novel verbs") Passives (full passives in *English* aren't produced until 4-5 years). Complexity of Constructiosn is increased by combinging simpler constructions "Look at Pete eating a bone" from "Look at XX" and "Eat XX" ``It is possible that the most appropriate model to describe the developmental process is some kind of prototype model in which some verb island constructions are seen as more central members of an abstract construction than others, for example give as the prototype for the ditransitive construction" cites Taylor 1996. ``The major problem for constructivisit theories has always been to explain how children can begin with utterances containing only words and end up with productive grammatical constructions based on abstract lexical and relational categories. Some theorists have thought the problem so serious that only theories that posit abstract structures as innate can ever be sufficient (Gleitman and Wanner 1982). But constructivist theries are able to account for development at least in principle if not yet in full detail, if the endpoint is characterized not in terms of the mathematics of formal grammars, but rather in terms of an inventory of linguistic symbols, categories, schemas and constructions--as in the work of Cognitive and Functional linguists such as Langacker, Fillmore, and Goldberg." pg 32. [my note to MIke: Thanks a lot for the clarification (about your article). The experiment that Ahktar is doing is really interesting. I'm surprised that 2 year olds can learn that kind of verb specific word order without any problem. But they presumably can't do that for each verb coming in indefinitely, since then they would have learned a completely hodge podge grammar. (I'm assuming they can't because you don't find hodge podge languages). So your understanding is that they can learn a certain number of verbs like that, but that since they are ultimately trying to categorize verbs into classes, the verbs can't remain islands forever. So the idea of a critical mass is pretty important. It makes sense that children need to categorize by the time they were trying to learn "too many" different isolated patterns. That idea is supported by some of Liz Bates' work with Judith Goodman. They argue that syntactic proficiency is strongly correlated with vocabulary size, which makes sense if it is the increasing vocabulary size that is forcing the syntactic generalizations. The reasons you give for why it is difficult to categorize verbs into classes make sense. And you might add to that the fact that each formally distinct construction tends to have a family of related senses (some related only by metaphor), and that each verb tends to appear with a variety of different constructions. So the categorization is just plain difficult (even for the linguist-analyst). Anyway, your learning scheme makes a lot of sense to me. Thanks again. PS You did make it clear that nouns are generalized much earlier, and I can see that labels to referents are categorized right away (also by animals, I agree). ] He notes that theere are many studies that purport to show that children understand word order, but he notes that these use already known verbs, and the child may only have acquired the verb specific word order pattern, not the abstract cosntruction. *!*Tuggy, David. 1993. Ambiguity, Polysemy and Vaguess. Cognitive Linguistics 4-3. 273-290. ``Within the Cognitive Gramar framework (Langacker 1987) ambiguity and vagueness may be seen as occupying opposite ends of a continuum with polysemy in the middle." pg 273 Traditionally vague items: neighbor, aunt Traditionally: (Lyons 1977, Zwicky and Sadock 1975): ambiguity (homonymy): two lexemes polysemy: a single lexeme with different distinct senses vagueness: a lexeme with a single but non-specific sense Polysemy is a sort of halfway point between ambiguity and vagueness. Paint
painting a portrait in oils on canvas painting a landscap with watercollors on paper painting trompe l'oeil on the interior wall and floor of a house pairing a mural on the exterior wall of a public building painting a decorative border on an interior wall painting the walls of a room with a single color of paint for decorative purposes, but also to preserve them painting the exterior of a houe primarily to preserve it painting furniture painting a car with an air gun painting stripes on a parking lot or roadway by driving a paint-spraying machine applying makeup applying iodine or colored disinfectant with a swabbing motion Results on linguistic tests are ambivalent {I have been painting and so has Jane (ok for senses 1 and 2; senses 1 and 10 are clearly zeugmatic. As Geeraerts also notes, context can influence the acceptability: {When I'm painting I try to get the color on evenly, and so does Jane. Syntactic differences: Paint is a verb of creation in senses 1-4: {Paint me a portrait. Not in some other senses : {??Paint me a car/??Paint me a house. Sense 12 normally has the disinfectant in object position. {Joe painted iodine on the wound. $>$ Joe painted the wound with idodine. [Also, {She paints. with null object can only have subset of readings] There are also figurative uses: {Paint a gloomy picture of the prospects for peace. The senses are related, and some of the senses are unitable, but different dictionaries choose to unite them differently. ``We see what is common between getting food ready to eat and counselling a person about to be traumatized, nd that accounts for us calling both activities preparing; yet it feels like a pun when you hear on the radio advice on how ot prepare your turkey for THanksgiving--by breaking it to him gently ("Tom are you religious?")pg 278. Another problem for strict dichotomy between ambiguity and vagueness is semantic change. It is common for vague meanings to become separate over time. Lewis 1967: cumulatve small changes occured in the meanigs of sense leading to ambiguity for sentence. ``Well entrenched structures, ceteris paribus, are more salient than less-entrenched structures: ie. they occur more energetically. " pg 279 CONTEXT also affects salience (pg 284). schema: what two (or more) cognitive structures have
"...there will always be some schema uniting any two cognitive structures, even if it be only the superschema ENTITY (Langacker 1987: 198). Tuggy: One can speak of a "schema" uniting checkmate and pawn, consisting of a game of chess (Note 7) [also for metonymy] "A schema is distant from its elaboration when relatively many speicifications of the elaborations must be despecified to form teh schema, and close when relatively few must be despecified." pg 280 A:cat, B: weasel, C:Mammal that eats mice (C is closer to A, B) A:dog, B: stool, C:thing with legs found in human dwellings (C is
[He doesn't really want to claim these C's are cognitive real, does he?] Figure 2: the ambiguity-vagueness cline Figure 3: the cline showing extensionality of semantic structures Figure 4: cline showing extent of semantic specifications Speakers can "filter out" specifications below a certain level
PUNS: context enhancing the salience of two meanings *!*Vahedi, Mohammad-M. ms. 1995. A new
V notes that Persian CPs can undergo derivational processes and then notes the following kinds of examples: {yek zamin-e saxti-i xord-am \\ one ground-EZ severe-IDF collide-1S \\ I fell down severely. (Ali also gets this one) {unha 'estaqbal-e garmi az Amir kard-and \\ they welcome-ez warm from Amir did-3PL \\ They gave Amir a warm welcome. (Ali: this isn't the same thing..) {Mina az man ejaze gereft vali Mahtab na-gereft \\ Mina from me permission got-3S but Mahtab NEG got.3S \\ Mina got permission from me but Mahtab did not. (Ali: this isn't the same thing) {agar zamin-i ke man xord-am to mixord-i mi-mord-i \\ if ground-IDF that I hit-1S you IND-hit-2S IND-die-2S \\ If you fell down the way I did, you would die. (Ali: didn't like it On the basis of examples like these, V argues that the host is a maximal projection (XP). He argues that they are syntactic products and are formed in the syntax. Szabolsci (1984): bleached existential predicates (contain primary logical components: EXIST, BECOME, BRING into EXISTENCE< DO, MAKE, GO, COME to BE, BECOME. Their non-logical semantic substance/content is "either totally bleached or backgrounded. They, however, keep an impoverished argument structure with an obligatory internal complement subcategorization frame." (pg 5) He claims that a weak or existential DO leads to the formation of CP (food-eat). Partee (1987) is cited for the claim that each syntactic category has a basic semantic type but can undergo type shifting. Substantiation: \begin{quote Every predicate of natural language must have some non-logical content. Therefore: If the meaning of a predicate contains at most logical constants and variables, it must enter into a "closest possible" syntactic realization with something whose meaning (also) contains some non-logical constant (Szabolsci 1986: 11). {quote To account for lexical properties, V assumes the theory of Lexical Relational Structure of Hale and Keyser (1988) H & K make Engish look like Basque in old Generative Semantics terms. Dear Mohammad, You cite Sproat (1988) as claiming that words shoudl be anaphorica islands, but Ward and Sproat and McKoon 1991 (Language) argue that words are not always islands: whether they are depends on pragmatic considerations. In English, Who gave Mary the hard time? is fine (felicitous in the situation in whcih you know someone gave Mary a hard time). Do you want to say that any indefinite object is a case of a complex predicate, whether it is marked with -i or bare? The "Substantiation" motivation for CPs you cite would seem to predict that the light verbs cannot occur alone (say with definite objects). How do you avoid claiming that? As an ex-student of George Lakoff's I don't know whether to smile or wince at Hale and Keyser's reinvention of abstract sytnax style analyses. *!*
Grammar account of Pronominal Anaphora Constraints. Language 71 2: 310-340. Basic data: Lakoff 1976: \eenumsentence{ *He loves John's mother. His mother loves John. Near him, Dan saw a snake. *Near Dan, he saw a snake. Reinhart's c-command: if the first branching node dominating the pronoun within the syntactic tree also dominates the full NP, coreference is ruled out. ``A node A c-commands node B if the branching node $\alpha$1 most immediately dominating A either dominates B or is dominated by a node $\alpha$2 which dominates B and $\alpha$2 is of the same category type as $\alpha$1.'' (Reinhart 1983:41). Problems for c-command: \eenumsentence{ *I spoke to him$_i$ about finances in Ben's office. *John gave a book to her for Sally's birthday. \\ (in the above, the pronoun does not c-command the NP, unless the PP node somehow doesn't count) *He went home and then Peter took a nap. He went into the study. \#John picked up the phone. (311)\\ (Reinhart: these are ruled out by discourse principles) *Her wish is that Sally's daughter will become a physicist(pg 312)\\ (again prounoun doesn't c-comand NP Sally See Kuno 1987, zribi-Hertz 1989 and Pollard & Sag 1992 for analyses of discourse principles, but assumption that discourse effects are distinct from syntactic principles which deal with "core" grammar. \newpage
! ACCESSIBILITY: pronouns ,reflexives, full NPs are used to signal differences in the accessililty or retrievability of their referents in a given context (Ariel 1988, 1990; Giv\'on 1989; Ward et al. 1991; Gundel et al. 1993). full NPs $<$ pronouns (scale of accessibility) \bigskip !
used to contextualize other elements (see Langacker 1991; 1993). Similar in the case of subjects to Chafe's (1976, 1992) and MacWhinney's (1977) ``starting points." \bigskip
! CONSTRAINT: ``A full NP (name or descriptive phrase) cannot appear in the dominion of a coreferential reference point, as this would conflict with the specification of a full NP as a low accessibilty marker." pg 314. \bigskip
in above examples is in a ``context" in which it should be highly accessible. This conflicts with the specification of the NP as NOT highly accessible. \bigskip Reference point organization is determined largely by semantic prominence. Two kinds of prominence: profiling and figure/ground asymmetry. \bigskip
! PROFILING: relevant for defining N, V, head, complement, modifier. \bigskip
! FIGURE/GROUND: basis for definition of grammatical relations. \bigskip
! BASE: the conceptual structures that an expression invokes that are crucial to its characterization. Roof invokes the structure of a building as its base; hypotenuse invokes the notion of a right triangle as its base. ($\approx$ frame) \bigskip
! PROFILE: the subpart of the conception which the expression designates. N's profile "things"; V, A, P's profile "relations"= interconnections between entities. V's are more specifically "processes," having temporal properties \bigskip
! FIGURE/GROUND: above/below profile same relation,but differ in choice for figure. \bigskip
! TRAJECTOR:Figure within a profiled relation. \bigskip
! SUBJECT: trajector of a profiled process (1987a:231-234). (this explains its tendency to correlate with discourse topic.) \bigskip
! HEAD: profile determinant \subsection{Complements vs modifiers One structure, D, is ! DEPENDENT on another, A, to the extend that
E.g., ON is dependent on THE TABLE in ``on the table" (THE TABLE is not dependent on ON). When the head is dependent, the autonomous element is a complement. When the head is autonomous, the dependent element is a modifier
Example: MAN ON THE MOON: MAN is head of fully NP, and is autonomous: ON THE MOON is a modifier. Within ON THE MOON, ON is dependent and THE MOON is a complement. (adapted from pg 319). ! Principles: (pg 320) R tends to be construed as a reference point of N to
R tends to be construed as a reference point of N if R precedes N in the linear string. (weaker than conceptual connectivity, most noticible when R and N are only weakly connected) \noindent Example of ! [1]: grammatical relations hierarchy: Subj $>$ DO $>$ Obl (Keenan & Comrie 1977) Interpreted in CG as prominence relations. The subj is a reference point with the rest of the clause in its dominion. (clause-internal topic: Giv\'on 1979a, Chafe 1976) DO is a reference point with all complements besides the subj in its dominion. (secondary figure Giv\'on 1984a, Langacker 1991). All complements of the verb are profiled within the clausal conception (modifiers/adjuncts are not profiled: only optionally construed as reference points). \bigskip
! [2]: making mental contact with one entity (bringing it into consciousness) can lead to mental contact with another entity which elaborates one of its substructures. \bigskip
[1] and [2] characterize a set of reference point configurations: COMPLEMENT CHAIN. The coreference phenomena addressed by c-command correspond to these chains. (pg 321) {*Jim put it$_i$ in the kitten$_i$'s box. pg 322 Ditransitive: first object is reference point with the second object
Only profiled nominal conceptions are so prominent as to necessarily be construed as reference points for other elements in the sentence. Other nominal conceptions are only optionally construed as reference points: \eenumsentence{ *He loves John's mother His mother loves John. John's mother loves him. In \ex{0a, he is necessarily a reference point and so (a) is bad. In \ex{0b, his is part of a modifer, not profiled and not necessarily a reference point. [question: isn't john in b a reference point for his? That is, the NP seems to be a reference point in the sense of being an element which is crucially needed in order to contextualize the pronoun. How is that possible?] \ex{0c shows that the possessive nominal can optionally act as a reference point. The same explanation explains why the following is possible: {After it$_i$ fell off the table, the ball$_i$ rolled across the floor pg 323 (The pronoun is embedded within a modifier and so is not profiled at the highest level of clausal organization). \bigskip
! The above can be captured by c-comand. The reason for the correspondence: c-command is defined so as to incorporate certain very schematic facets of conceptual organization. The subject is attached to the highest node; the DO is attached directly to a major constituent. \bigskip
! [3] Linear Order: \eenumsentence{ John checked the mailbox. There was a package for him. He checked the mailbox. \#There was a package for John. (pg 325 [note He$_{jm$ looked in vain for a package addressed to John Moore would be possible in a story.-AG] \eenumsentence{ ??Near Dan, he saw a snake $>$ *He saw a snake near Dan. ! Process-internal modifiers: Strongly interconnected with the subj (such modifiers are construed as within the subject's dominion, regardless of word order) non-argument instruments, goals, sources spatial, temporal scene-setters modifiers whcih further characterize participants (e.g. characterize the mental state of the agent) \eenumsentence{ Sally nudged Sam with her umbrella. *With Sally's umbrella, she nudged Sam. pg 327 [Come up with examples of others] \medskip ! Process-external modifiers: Weakly interconnected with the subj (Linear order plays a strong role in whether modifier is considered within the dominion of the subject) space builders (Fauconnier 1985) modifiers which relate the clause as a whole to the larger discourse Afterthoughts (bolinger 1979): comments on the clause as a whole \eenumsentence{ In Kathleen Turner's latest movie, she falls in love with Tom Cruise (pg 327) He lied to me -- something that John was rather fond of doing. (from Bolinger 1979: 298) DO is strongly connected only to those elements with which tải về 302.24 Kb. Chia sẻ với bạn bè của bạn:
|