Introduction to Contemporary Linguistics November 4, 1998 Semantics: the meaning of words, and the meaning of sentences OVERVIEW: 1. Word meaning 2. Sentence meaning 3. Discourse and pragmatics 4. Summary ============================================================= 1. Word meaning >What does a word like "book" or "®Ñ" mean? There are two meanings of "meaning": >"Book" has a referent: something in the real world: (actually, this is just a picture....) >"Book" has semantic properties: mental units of meaning: "concrete" (not abstract) "inanimate" (not animate = not alive) "artificial" (made by people, not part of nature) >Semantic properties exist in people's minds, not necessarily in the real world. >"Cold" and "hot" seem related because they share semantic properties: cold hot "temperature" "temperature" "low" "high" >However, "cold" and "hot" have totally unrelated referents in the real world! >Semantic properties are crucial for understanding how human beings understand meaning in language. >The more semantic properties shared by two words, the more similar their meanings seem to be: child cat flower living living living animal animal plant human nonhuman nonhuman >Often the semantic properties of adjectives and verbs have a kind of constituent structure: dead = not [ living ] die = become [ not [ living ] ] kill = cause [ become [ not [ living ] ] ] >Semantic properties help us understand the changes that derivational morphology makes: simple + ify = cause [ become [ simple ] ] >Semantic properties also play a role in choosing noun classifiers (¤ÀÃþµü) in Chinese: Classifier Semantic properties Examples ±i flat, broad ®à¤l¡B¯È ±ø long, narrow ¸ô¡B½u¡B³D¡B³½ °¦ animal ª¯¡BµU¤l Áû round, small ¤ú¾¦¡B¸²µå ­Ó people, abstractions, ¤H¡B¹Ú¡B¦è¥ other objects, ...? ®à¤l.....¡H >Synonyms (syn + nym = "same" + "name"): words that share almost all semantic properties. Perfect synonyms do not exist: depart vs. leave (long time vs. any length of time) >Antonyms (ant + nym = "opposite" + "name"): words that are opposite in meaning. But what does "opposite" mean? dead vs. alive (complementary: "dead" = "not alive") small vs. big (gradable: a small elephant is bigger than a big mouse) buy vs. sell (relational: if A buys from B, then B sells to A) >If something has more than one meaning, then it is ambiguous. Lexical ambiguity can happen in two different ways: >There might actually be two historically unrelated morphemes that just happen to sound (or look) the same: homonyms (homo + nym = "same" + "name"). [Homophones sound the same, and homographs are written the same.] (1) a. bat b. [thu]: "two", "to" and "too" c. ³Q: preposition vs. noun (³Q¤l) d. [i1]: ¤@¡Aå¡A¦ç¡A¨Ì¡Aäv¡A¥ì¡A... >There might really be just one morpheme, but its single meaning has split into related meanings: it is polysemous (poly + sem = "multiple" + "meaning"). (2) paint: kind of liquid (N); using this liquid (V) (3) grow = become [ big ] ("The tree grew.") grow = cause [ become [ big] ] ("The tree grew leaves.") >Do you think these examples show different homonyms, or just one morpheme that is polysemous? (4) a. ¥d¦í (5) a. ¶}ªù b. ¥d¨® b. ¶}¿O c. ¥d¤ù c. ¶}¨® d. ¥d³q d. ¶}ª±¯º 2. Sentence meaning >Sentences mean more than just the meanings of their words: (6) a. The dog bit the man. b. The man bit the dog. >These sentences have the same words, but the syntax makes a semantic difference: The syntax determines the semantic roles of the subjects and the objects. >The two major semantic roles (also called thematic roles): >The agent is the one that actively does something. >The patient is the one that is passively affected. >The above sentences are active, so the syntactic subject is the semantic agent, and the syntactic object is the semantic patient: The dog bit the man SYNTAX: subject object (of V) SEMANTICS: agent patient >When you use the passive transformation, these semantic roles switch places: The man was bitten by the dog SYNTAX: subject object (of P) SEMANTICS: patient agent >Verbs can differ in the semantic roles they use, and this information must be put in their subcategorization: (7) The dog chased the man agent patient (8) The dog hated the man experiencer theme >Experiencer is a role for something that doesn't do an action, but just experiences an emotion. >Theme is a role for something related to the action, but not directly affected by it. >This subcategorization has interesting effects on syntax! >In a sentence, each NP can only have one semantic role: (9) a. The dog chased and bit the man. b. ?*The dog hated and bit the man. (21b sounds bad since "hate" and "bite" require different semantic roles) >Because the passive transformation makes the patient into the subject, only verbs that subcategorize for a patient can be passivized: (10) Her old dog resembles ET theme theme (not patient!) (11) *ET was resembled by her old dog. >English verbs that make the subject a theme rather than an agent are confusing to Chinese students. (12) World War III will happen. theme (not agent) >Syntax can interact with semantics in other ways, too: >Structurally ambiguous sentences (remember?) >The interpretation of pronouns: (13) a. John1 hates himself1. ("himself" = "John") b. *John1 hates himself2. ("himself" ¹ "John") (14) a. John1 hates him2. ("him" ¹ "John") b. *John1 hates him1. ("him" = "John") >Himself, herself, etc are reflexive pronouns: their meaning "reflects" back to some NP in the sentence. (15) a. Bill1 thinks that John2 hates himself2. b. *Bill1 thinks that John2 hates himself1. (16) a. Bill1 thinks that John2 hates him1. b. *Bill1 thinks that John2 hates him2. >There is a very simple pattern here: [OVERHEAD] >reflexives (like himself) must refer back to an NP inside the same clause (S) >ordinary pronouns (like him) must refer back to an NP outside the clause (S) >The effect of definiteness: how specific or unique an NP is. >"There" can only be used with indefinite NP's: (17) a. There is a stranger in my room. b. *There is the stranger in my room. c. *There is James in my room. >Subjects in Mandarin are (usually) definite; otherwise you need to use "¦³": (18) a. ¨º­Ó¤H¨Ó¤F¡C b. ?¤@­Ó¤H¨Ó¤F¡C c. *¦³¨º­Ó¤H¨Ó¤F¡C d. ¦³¤@­Ó¤H¨Ó¤F¡C >Sentence meaning also interacts with inflection (morphology that marks syntactic information). >In English, passive syntax/semantics requires passive inflection: (19) a. The book was stolen. b. *The book was stole. >In Mandarin, ¤F [le] represents two distinct homonyms: >When used at the end of a sentence, ¤F usually represents a sentence-final pragmatic particle (we'll discuss this in the next section). >However, when used right after a verb, ¤F usually represents an inflectional suffix. >The inflectional suffix ¤F does not mark tense. Instead, it marks another time property called aspect. >Tense marks the relationship between the time of an event and the time of the sentence ("now"): (20) a. I danced. [past tense] b. I dance. [present tense] c. I will dance. [future tense] >Aspect indicates the internal time structure of the event itself, independent of the time of the sentence. English marks aspect separately from tense: (21) I am dancing. [progressive aspect, present tense: dancing continues now] (22) I was dancing. [progressive aspect, past tense: dancing continued before now] (23) I have danced. [perfective aspect, present tense: dancing is limited in time, seen from now] (24) I had danced. [perfective aspect, past tense: dancing is limited in time, seen from before now] (25) I will have danced. [perfective aspect, future tense: dancing is limited in time, seen from after now] >The verb suffix ¤F marks perfective aspect, which is used for events that are limited in time, with a definite start and end. >Even without ¤F, the action can be in the past, but the past event doesn't have to be limited to some specific time: (26) a. §Ú¼g¨º¥»®Ñ¡C b. §Ú¼g¤F¨º¥»®Ñ c. ¨º¥»®Ñ¬O§Ú¼gªº¡C d. *¨º¥»®Ñ¬O§Ú¼g¤Fªº¡C >Events can be limited in time even if they're not in the past, so you can use ¤F even when it's not past tense: (27) ©ú¤Ñ§Ú´N¶}°£¤F¥L¡I (28) §Ú¦Y§¹¤F§A¤~¦Y¡C >In a moment I'll discuss the sentence-final ¤F . >Even though syntax and semantics interact, don't forget that they are distinct parts of the grammar!!!! >Syntactic grammaticality without semantic content: (29) Colorless green ideas sleep furiously. >Semantic content without syntactic grammaticality: (30) *Brown little dogs eat noisily their food. ("Little brown dogs eat their food noisily.") 3. Discourse and pragmatics >Discourse: the unit of language made up of sentences (for example, a lecture, a conversation, a book): morphemes < words < sentences < discourse >Some words only have meaning in the context of a discourse: (31) I read a book. [book not yet mentioned in the discourse] (32) I read the book. [definite book that has already been mentioned in the discourse] >Pragmatics: the effect of real-life context on language. Literal meaning Pragmatic use (33) It's too hot in here! Turn on the AC! (34) I'm hungry! Give me some food! >Pragmatics includes presuppositions, the unspoken assumptions shared by speaker and listener: (35) Would you like another beer? (presupposes that you already had some) (36) Have you stopped kicking dogs? (presupposes that you used to kick them) (37) It was Mary's birthday. John gave her a sweater. (people give presents, like sweaters, for birthdays) >Another part of pragmatics is deixis (N [DAJKsis]): words or expressions are deictic(Adj [DAJKtik]) if they change meaning depending on the context in which they are used: "I" = "James" only if James uses it "you" = "James" only if someone is talking to James "this" = the definite thing near the speaker "that" = the definite thing not near the speaker >In Mandarin, all definite determiners are deictic: ¨º¡B³o English has a definite determiner that is not deictic: "the" >By the way, many jokes play with pragmatics [OVERHEAD] >Unlike English, Mandarin has many sentence-final particles (little function words) that mark pragmatics: §a¡A©O¡A°Ú¡A³á¡A¤F... >For example, what does °Ú mean? In a sense, nothing. Instead, it has a pragmatic function: it "softens" the message, which may make it more friendly: (38) a. ¤p©ú¡I (39) a. ¦Y¶º¡I b. ¤p©ú°Ú¡I b. ¦Y¶º°Ú¡I >What does sentence-final ¤F do? Most scholars agree that it has the pragmatic function of indicating a currently relevant state, which usually means a state that is relevant right now to the listener. Unlike the other morpheme ¤F , this has nothing to do with the aspect of an event! (40) a. ¤U­Ó¤ë§Ú´N¦b¤é¥»¡I [just a fact] b. ¤U­Ó¤ë§Ú´N¦b¤é¥»¤F¡I [... and you should know this because maybe you were planning to visit me] (41) a. §Ú³Ü¤F¤TªM©@°Ø¡C [fact] b. §Ú³Ü¤F¤TªM©@°Ø¤F¡C [... so don't give me more!] (42) a. ¦oª¾¹D¨º­Ó®ø®§¡C [fact] b. ¦oª¾¹D¨º­Ó®ø®§¤F¡C [... and she didn't know before] >Here is an excellent description of Mandarin sentence-final particles (and aspect, and passives, and many many other parts of Mandarin morphology, syntax and semantics): Li, C. N. and Thompson, S. A. (1981) Mandarin Chinese: A Functional Reference Grammar. University of California Press. [also available in Chinese translation] 4. Summary >A semantic comparison of English and Mandarin: [this table is messed up as usual] English Mandarin Homophones some: bat, two/to/too, write/right, ... many, many, many: ¤@¡Aå¡A¦ç¡A¨Ì¡Aäv ... Noun classifiers none (?) many: ±i¡A±ø¡AÁû¡A°¦¡AªK ... Indefinite subjects Common Less common Deixis Not all definite markers are deictic: "the" All definite markers are deictic: ¨º¡A³o Sentence-final pragmatic particles none (?) several: §a¡A©O¡A°Ú¡A³á¡A¤F ...