A contextual approach to lexical semantics n n n n n n npřednáška nThe approach described is a variety of 'contextual' approach' . It is assumed that the semantic properties of a lexical item are fully reflected in appropriate aspects of the relations it contracts with actual and potential contexts. nIn theory, the relevant contexts could include extra-linguistic situational contexts.. nThere are good reasons for a principled limitation to linguistic contexts: 1.the relation between a lexical item and extra-linguistic contexts is often crucially mediated by the purely linguistic contexts 2.Any aspect of an extra-linguistic context can in principle be mirrored linguistically. 3.linguistic context is more easily controlled and manipulated nWe should seek to derive information about a word's meaning n from its relations with actual and potential linguistic contexts nCharacteristics of words in utterances are n constrained not only by their meanings, but also by their grammatical properties. nGrammatical constraints may overlap and reinforce semantic ones, but they may also be semantically arbitrary. Meaning and grammar ndistinction - is not an easy task nThey are intimately interwoven nThe distinction between grammar and meaning has a strong intuitive basis. n deviation nSemantic deviance nHe harvested a magnetic puff of amnesia. n nDeviance of grammar nThem yesterday goed to home. n n nIt's too light for me to lift. n nI've nearly completed. - (in answer to How are you getting on with those jobs I asked you to do?) n nthe deviance disappears completely if light is substituted n by the semantically distinct, but syntactically identical, heavy nthe deviance can be cured by inserting them after completed. n nsyntactic deviances can be readily corrected,whereas semantic deviances cannot. For example: nHe harvested a magnetic puff of amnesia. n nThem yesterday goed to home. n n nMore promising strategy is to ask not how or whether a deviant sentence can be corrected, but what the minimal changes are that will render it normal, then we examine the nature of the changes. nIf a deviant sentence can be normalised by adjusting its grammatical structure - changing the order or syntactic category of elements, or by adding substituting or deleting one or more grammatical elements- GRAMMATICAL DEVIANCE nIf the minimal change required is one necessarily n involving one or more full lexical items- SEMANTIC DEVIANCE n nthis procedure would be more informative if we were able to characterise grammatical and lexical elements more explicitly. n nJohn's kindness amazed Mary. n nThis is, in fact, possible in terms of what are called nclosed set items (are those belonging to classes whose membership is virtually constant during the lifetime of an individual speaker. They have few or no possibilities of substitution in an actual sentence) naffixes (dislike, kindness , John's, waited, coming, blacken...) and independent words (sometimes called markers), such as articles,conjunctions, prepositions and so on. nmajor part of whose linguistic function is to signal the grammatical organisation of sentences nand open set items (are those which belong to classes which n are subject to a relatively rapid turnover in membership. They are the lexical roots -the principal meaning-bearing elements in a sentence. n n n nJohn's kindness amazed Mary nBill- cool- amuse- Sue nMary- rude- disturb- John nSue - sad- shock- Bill nIf the minimal change required to 'cure' an anomaly in a sentence involves one or more closed set items, then the deviance is grammatical. nIf, however, the sentence can most easily be normalised by replacing one or more open set elements, then the deviance is semantic. nHe harvested a magnetic puff of amnesia. nTo normalise, the lexical roots must be altered.-: He exhaled a carcinogenic puff of smoke. nThem yesterday goed to home nAll the changes needed to normalise involve closed set items. nIt is possible for a sentence to exhibit semantic and grammatical deviance simultaneously : n The green idea sleep. nTwo separate operations are needed to normalise this sentence: n1. one involving closed set items: nThe green idea is sleeping. n2. the other involves an open set item : n The green lizard is sleeping nIt is not possible to disentangle semantics from grammar completely. One reason for this is that many grammatical elements are themselves bearers of meaning - this is true, for instance, past tense affix -ed, and the plural affix –s. nBut otherwise the meaning they carry is not of a radically different sort from that carried by lexical roots, and grammatical and lexical elements frequently interact semantically. nI visited Arthur next week. nI shall visit Arthur next week. nI visited Arthur last week. n nSome sentences can be normalised either by grammatical adjustment or by lexical adjustment. nThe cake was baken. nThe cake was baked. nThe cake was taken. Criteria for deciding whether an anomalous sentence is semantically or grammatically deviant : nAn anomaly which can only be removed by replacing one or n more open set items is semantic; n n nan anomaly which cannot be removed by replacing one or more n open set items, but can be removed by changing one or more n closed set items, is purely grammatical; n nan anomaly which can be cured either by changing one or more n closed set items or by replacing one or more open set items n is semantic (albeit with grammatical implications) if the open n set replacements are distinguished by the possession of certain n semantic properties; otherwise, it is purely grammatical, nA poetic context can also condition the reader or hearer to accept grammatical deviance, especially if syntactic well-formedness is clearly being sacrificed to some higher aesthetic end, such as the maintenance of rhyme,or metre, or some other patterning The data of semantics nFor a study of lexical semantics, there would seem to be two principal sources of primary data: n nOne source is the productive output, spoken or written, of native users of the language. n nThe second principal source of primary data on which a study of lexical semantics can be based is furnished by intuitive semantic judgements by native speakers of linguistic materials of one kind or another. > nIt might seem obvious that, if one is studying word-meanings, one ought to find native speakers' intuitions concerning the meanings of words the most informative. nThe intuitions most relevant to a study of meaning would seen at first sight to be intuitions about what things mean. n nHe watched it with intense concentration for a few moments, n then left the room. n nHe looked at it with intense concentration for. a few moments, n then left the room. n nThe ability to 'explain' meanings is an uncommon skill. nThe answer is to elicit not intuitions OF meaning, but intuitions ABOUT meaning n DISCIPLINING INTUITIONS ●No empirical science can operate without human intuitive judgement intervening at some point. ●One of the simplest and most basic semantic judgements one can make concerning an utterance in one´s native language is whether it is to some degree odd or not. ●What can informants do is to distinguish a fully normal sentence from one which is to some degree odd. ●An odd sentence is not necesarilly meaningless, or incapable of conveying a message. On the contrary, an oddness of one sort or another is frequently a signal that an expression is being used creatively, in a novel extension of its usual sense. n ●ZEUGMA-inappropriate linking, e.g.Arthur and his driving licence expired last Thursday. ●PLEONASM, can be normalised by replacing one of its elements with something more specific ●A female mother. ODD ●A lesbian mother. NORMAL ●Kick it with one of your feet. ODD ●Kick it with your left foot. NORMAL ●DISSONANCE, can only be be cured by replacing one element by something less specific, conflict of words. ●The cat barked. ODD ●The animal barked. NORMAL ●IMPROBABILITY ●He walked by car. ●The sun very shines today. SEMANTIC ANOMALIES ENTAILMENT ●A proposition P is said to entail another proposition Q when the truth of Q is a logically necessary consequence of the truth of P (and the falsity of P is a logically necessary consequence of the falsity of Q). ●We shall use the term to refer to an analogous relation between sentences. ●A sentence THAT´S A DOG can be used to express an indefinetly large numbers of proposions ●The sentence THAT´S A DOG entails the sentence THAT´S AN ANIMAL and can be viewed as a kind of shorthand for the pattern of normality like the following: ●IT´S A DOG THEREFORE IT´S AN ANIMAL ●IT CAN´T POSSIBLY BE A DOG AND NOT BE AN ANIMAL. TYPES OF ENTAILMENT n1. UNILATERAL ENTAILMENT ●It´s a dog unilaterally entails It´s an animal. n2. MUTUAL ENTAILMENT ●The meeting began at 10.00 a.m. And is entailed by The meeting commenced at 10.00 a.m. n3. CONTRARIETY ●It´s a cat and It´s a dog stands in a contrary relation: It´s a cat unilaterally entails It´s not a dog. n4. CONTRADICTION ●It´s dead entails and is entailed by It´s not alive. INTUITIVE JUDGEMENT ●Another useful and reliable intuition is that of recurrence of semantic contrast or semantic proportion. ●For instance, speakers are well able to judge that the contrast between 1A and B is the same as that between 2 A and B. ●1A I like him ●1B I dislike him. ●2A They approved of the idea. ●2B They disapproved of the idea. THE MEANING OF A WORD ●We shall say, that the meaning of a word is fully reflected in its contextual relations and that the meaning of a word is constituted by its contextual relations. ●We can picture the meaning of a word as a pattern of affinities and disaffinities. ●A SYNTAGMATIC AFFINITY-there is a syntagmatic affinity between DOG and BARKED, since THE DOG BARKED is normal ●A SYNTAGMATIC DISAFFINITY is revealed by a syntagmatic abnormality, THE LIONS ARE CHIRRUPING. ●PARADIGMATIC AFFINITY- cat and dog have a fairly high degree of P.A.: ●I STROKED THE CAT/DOG. ●WE HAVE A CAT/DOG AT HOME. ● SEMANTIC TRAIT ●A particular word-meaning which participates in this way in the meaning of another word is termed a SEMANTIC TRAIT of the second word. ●We shal define a number of statuses of semantic traits: ●CRITERIAL, EXPECTED, POSSIBLE, UNEXPECTED, EXCLUDED. SEMANTIC TRAITS ●CRITERIAL and EXCLUDED traits can be diagnosed by means of entailment relations between sentences. ●For example, „animal“ is a criterial trait of dog because IT´S A DOG entails IT´S AN ANIMAL ●„FISH“ is an excluded trait of dog because IT´S A DOG entails IT´S NOT A FISH. ●For the diagnosis of expected, possible and unexpected traits, the but-test is extremely useful. EXPECTED, UNEXPECTED, POSSIBLE TRAITS ●Consider the status of „can bark“ as a trait of dog, „can bark“ is not a criterial trait. ●EXPECTED TRAIT: ●It´s a dog, but it can bark. (ODD) ●It´s a dog, but it can´t bark. (NORMAL) ●The sort of oddness may be termed EXPRESSIVE PARADOX. ●UNEXPECTED TRAIT: ●It ´s a dog but it can sing. (normal sentence, unusual dog, „can sing“ is unexpected trait of dog ) ●It´s a dog, but it can´t sing. (EXPRESSIVE PARADOX) ●POSSIBLE TRAIT: ●A possible trait is signalled when both test sentences exhibit expressive paradox, and P and Q is normal. ●It´s a dog, but it´s brown. ●It´s a dog and it´s brown. ● SEMANTIC TRAITS-TASK ●TASK: ●Consider the relation between „adapted for flight“ as a semantic trait of bird, and „possesses four legs“ as a trait of a dog. They are alike in that neither is criterial, both are expected. ●It´s a bird does not entail IT IS ADAPTED FOR FLIGHT. (There are birds such as ostrich, kiwi..) ●It´s a dog does not entail IT HAS FOUR LEGS. (A dog may have a birth abnormality...) ●Create 2 sentences with BUT, 1 would be normal, 2 is odd: TASK-SOLUTION ●It´s a bird, but it´s adapted for flight. (odd) ●It´s a bird, but it´s not adapted for flight. (normal) ●It´s a dog, but it has four legs. (odd) ●It´s a dog, but doesn´t have four legs. (normal) CANONICAL TRAITS ●Semantic traits whose absence is regarded as a defect will be called CANONICAL TRAITS. ●Canonical traits can be distinguished from non-canonical expected traits in a number of ways: ●The typical bird is adapted for flight. ●Birds are typically adapted for flight. ●What kinds of bird are not adapted for flight? ●Canonical traits are not only to be found in words denoting living things.