1	Pragmatics and Discourse (Doing things with language)
2	Pragmatics Situation-based use of language The study of language use: The way language is used in communicative situations The way we interpret language utterances from context How we understand non-literal expressions How we plan and execute utterances to fit expectations, intentions, context
3	Implications and entailment What the speaker/writer means or implies (vs. what s/he literally says) English: kill = cause + become + dead Conversational implicature: discourse obligations, (im)plausible conclusions Philosophy, formal logic has much to contribute to this area
4	Logical inferences Modus Ponens: p à q p -------- q Modus Tollens: p à q Ø q --------- Ø p Hypothetical syllogism: p à q q à r -------- p à r Disjunctive syllogism: p v q Ø p -------- q
5	Formal logic and inferences DeMorgan’s Laws Ø(j v y) ó (Øj & Øy) Ø(j & y) ó (Øj v Øy) Conditional Laws (j ® y) ó (Øj v y) (j ® y) ó (Øy ® Øj) (j ® y) ó Ø (j & Øy) Biconditional Laws (j « y) ó (j ® y) & (y ® j) (j « y) ó (Øj & Øy) v (j & y)
6	Presupposition John regrets that he ate all the eggs. John is sorry that he ate all the eggs. John repents of having eaten all the eggs. John is unhappy that he ate all the eggs. John feels contrite about eating all the eggs. John feels penitent about eating all the eggs. John feels remorse for having eaten all the eggs. all presuppose: John ate all the eggs.
7	Coreference Behind [Mary]_i [she]_i heard the snake. [Herself]_i is proud of [her]_i. If [that jerk]_i calls, tell [Tom]_i I’m busy tonight. [He]_i insists that [the electrician]_i found nothing wrong. [Mary]_i told [John]_j that [they]_/j/i/k were assigned clean-up duty.
8	Referentiality The man drinking Postum is Fred. The man who can lift this stone is stronger than an ox.
9	Referent resolution Syntactic coreference principles (Binding Principles A-C, φ-features, distance) The Hobbs algorithm: ordering of syntactically-derived potential antecedents from previous utterances The centering algorithm: cache of 2-3 previous sentences, stack-based memory, preference weighting among possible antecedents
10	Dialogue and discourse Discourse processing: interpreting language in its context Deixis: pointing to real-world entities Conversation and situation Beliefs, desires, intentions Turn-taking
11	Discourse fragments A: John’s cooking tonight. B: Where’s the Alka-seltzer? I now pronounce you man and wife. A: Do you know what time it is? B: Yes. A: That was a great movie. B: You can say that again! A: That was a great movie.
12	Speech acts Austin: “How to do Things with Words” Huge topic Psycholinguists, philosophers, anthro-pologists, literary critics, lawyers, linguists While saying something, or after having said it, we DO (accomplish) something with our utterance; “performatives”
13	Example speech acts I bet you six dollars it will snow tomorrow. I hereby christen this ship the USS Clinton. I apologize. I therefore sentence you to 5 years hard labor. I bequeath you my 1830 edition of the BoM. I declare war on Iraq. I give you my word. I pledge my allegiance… *I hereby divorce you.
14	Speech acts (cont.) Felicity conditions (else misfire [A,B], abuse [B]) Must exist a conventional procedure, effects; situation and persons must follow it Procedure must be executed correctly and completely Participants must have requisite thoughts, feelings, intentions; subsequent conduct must be followed
15	Assumptions and obligations Preparatory conditions S is able to perform A (May I breathe for you?) H wants S to perform A (May I kill your cat now?) Sincerity conditions (S intends A) Propositional conditions S1 promise A è S1 achieve A S1 request A è S2 address R; accept/reject A S1 YNQ whether P è S2 answer-if P Essential conditions (S obliged to do A)
16	Act types Locutionary act: uttering of the sentence Illocutionary act: whether making promise, offer, statement, etc. Perlocutionary effect: effects on hearer(s) Shoot him! It’s cold in here. Do you know what time it is?
17	Selecting speech acts Close the door. I want you to close the door. I’d be much obliged if you’d close the door. Can you close the door? Are you able by chance to close the door? Would you close the door? Won’t you close the door? Would you mind closing the door? Would you be willing to close the door? You ought to close the door. It might help to close the door. Hadn’t you better close the door? May I ask you to close the door? Would you mind awfully if I were to ask you to close the door? I am sorry to have to tell you to please close the door. Did you forget the door? Do us a favor with the door, chap. How about a bit less breeze? OK, Johnny, what do big people do then they come in?
18	Conversation maxims (Grice) Quantity: be sufficiently verbose Not more/less informative than necessary Quality: be truthful Don’t say what you know is false or unsupported Relevance: be relevant Say only things that are relevant Manner: be perspicuous Say things unambiguously, clearly, briefly, orderly
19	Flouting the maxims Purposely violating conventions; carries conversational implicature Flouting relevance: A: Would you like some fresh brownies? B: Is the pope Catholic? Flouting quantity: A: What are you reading? B: A book. Flouting quality: Your are the cream in my Postum… Flouting manner: Il m’ennuie, celui-là…
20	Basic discourse production
21	Deixis Pointing, indicating: context is indispensable Meet me here a week from now with a stick this big. Context: Speaker(s), addressee(s), utt-time, utt-place, indicated objects, shared assumptions/knowledge
22	Deictic types Time: (I’ll be back in an hour. It’s now 12:13. There is a man on the moon. This afternoon the Dow rose 24 points.) Place: (This city is really beautiful. The cat is behind the box. This side of the box is red.) Person: (He’s not the duke, he is; he’s the butler. I’m Deryle. Are you French?) Discourse: (Anyway, … / In conclusion, … / …the figure on the right. / Still, … / That was the funniest story I ever heard.)
23	Coherence Network of conceptual relations that underlie a text/conversation (Cf. cohesion: surface relations) A subjective property of texts (Cf. cohesion: objective) Dependent on expectations and real-world experience, cultural background
24	Centering theory Discourse referents, zero anaphors Metric for ranking possibilities Very useful in pro-drop languages Referent tracking, anaphora
25	DRT Represent, track discourse entities (explicit and implicit) Predicate logic for deduction, inferencing, etc. Used in many dialogue, discourse analysis settings
26	Sample DRS Mary left on January 1^st. n=now (utterance time) t=reference time t’=event time x=discourse entity
27	Another sample DRS Since she arrived, Mary has been busy.
28	Dialogue analysis, design Analyzing dialogues Turn-taking, linguistic content, pauses Appointment scheduling, conference registration, travel planning (airlines) Annotating corpora Use in conversational systems (TRAINS, JANUS, VERBMOBIL) Active area in e-business, chatterbots
29	TRAINS dialogue utt1 : s: hello <sil> can I help you utt2 : u: yeah I want t- I want to determine the maximum number of boxcars of oranges <sil> that I can get to Bath <sil> by seven a.m. <sil> tomorrow morning utt3 : so <brth> hm <sil> so I guess all the boxcars will have to go through oran- <sil> through Corning because that's where the orange juice <brth> orange factory is utt4 : so from Corning to Bath how far is that utt5 : s: two hours utt6 : u: and it's gonna take us also an hour to load <sil> boxcars right utt7 : s: right + + utt8 : u: + okay + so <sil> hm so <sil> every trip will take at least <sil> three hours <sil> then utt9 : um utt10 : s: right we can unload any amount of cargo onto a train in one hour utt11 : so we can + <sil> do a maximum of three + boxcars in an hour utt12 : u: + right <sil> okay + utt13 : okay <sil> so I guess one thing we can do oh <brth> so <brth> I guess one thing is that we should see how many boxcars we can actually get to Corning in four hours utt14 : um how far is it from Avon to Bath <sil> to Corning utt15 : s: <click> <brth> that's six hours it's + shorter + through Dansville utt16 : u: + okay +
30	Scripts E.g. action BUY(e): Roles: buyer, seller, object, money Constraints: Human(buyer), SalesAgent(seller), IsObject(object), Value(money, Price(object)) Preconditions: AT(buyer, Loc(seller)), OWNS(buyer,money), Owns(seller,object) Effects: ~Owns(buyer,money), ~Owns(seller,object), Owns(buyer,object), Owns(seller,money) Decomposition: GIVE(buyer,seller,money), GIVE(seller,buyer,object)
31	Other issues Implication, entailment, presupposition Thematic presentation strategies Serialization strategies (addresses, lists of languages, etc.) Socio-cultural norms for language use (forms of address, taboo, etc.) Motivated deviation from pragmatic norms
32	Chatterbots Computer programs (usually agents) that hold conversations with people Most limited to specific domains (e.g. Hampy) The Loebner contest “An average interrogator will not have more than a 70% chance of making the right identification after five minutes of questioning” (1950) 6 c-bots, 4 humans (teacher, MBA, minister, yogi), 10 judges (linguist, chemist, 2 phil’s, musician, psych’ist, journalist, u’grad, author, grad) 91% correct after 5 minutes, 93% after 15 No computer was mistaken for a human.
33	Sample bot chatter A: I like techno, especially retro-electronic like Kraftwerk. I like 70's bands like ABBA. Also I like Opera. J: ALICE LIKES ABBA A: That's me. My favorite group is Kraftwerk. J: FIGURES A: No but I know a lot of female humans. J: DO YOU SPEAK ANY SWEDISH A: I speak only English right now, but I'd like to learn some other languages. Can you teach me a few phrases in Any Swedish. J: SURE A: OK let's do it. Tell me more. J: SWEDISH FOR COMPUTOR IS DATOR A: How do you usually introduce yourself? J: WOMAN IS KVINNA SO I GUESS YOU ARE A KVINNLIG DATOR A: What do you look like? J: I LOOK TONS DIFFERENT FROM YOU A: You sound very attractive. J: HOW WOULD YOU DESCRIBE YOURSELF BY THE WAY A: It is a good question. Let me think about it. I can ask someone about it.