π Lecture 31 β POLITENESS AND CONTEXT
π Overview: This lecture explores politeness as a pragmatic phenomenon where meaning is determined not by linguistic form but by social function and context. It examines how situational, social, and cultural factors influence the choice of politeness strategies, and applies these concepts to real-world examples of requesting and conversational interaction.
ποΈ Topics Covered
The lecture covers five main topics: Form and Function (distinguishing politeness from deference and polite forms from polite intentions); Situational Context (size of imposition and formality as factors); Social Context (social distance, power relations, and demographic variables); Cultural Context (cross-cultural variation in politeness norms); and Applying Politeness (practical analysis of requesting strategies and dialogue using politeness theory).
π Lecture Summary
Topic β 155 Form and Function
Politeness is a pragmatic phenomenon; it lies not in the form and the words themselves, but in their function and the intended social meaning. Consider the example: "If you'd be as kind as to shut up, I'd appreciate it" (Elmore: Hombre, 1989). Here the form is polite but the intention is not. If a speaker uses more polite forms than the context requires, the hearer might suspect there is an intention other than redressing a face-threatening act (FTA).
Another example of inappropriate use of polite forms is a man's request to his pet: "Cat, I wonder if you could possibly let me have my seat back?" β meant to entertain whoever hears it. Politeness is not the same as deference, which is a polite form expressing distance from and respect for people of a higher status. Deference is rare to find grammatically signaled in English, although it is present in honorifics such as 'Sir' and 'Madam'.
π Definition β Face-Threatening Act (FTA): An utterance that potentially threatens the positive or negative face of the hearer (or speaker). π Definition β Deference: A polite form expressing distance from and respect for people of a higher status (e.g., 'Sir', 'Madam').
Topic β 156 Situational Context
Politeness is influenced by the elements of the context. There are two situational context factors that influence the way we make a request. One is the size of imposition β the routines and the reasonableness of the task. The rule is: 'the greater the imposition, the more indirect the language is'. For example, to borrow a large amount of money (Rs.50,000), one might employ a series of hedges and other negative politeness phenomena: "I couldn't borrow Rs.50,000/-, could I, if you don't need it right now?" On the other hand, to borrow a small sum (Rs.10), a request could be bald on record: "Give me 10 rupees."
The second factor is the formality of the context. The tendency is: 'the greater the formality, the more indirect language is'. A student sitting informally in the common room over coffee might stop a colleague from interrupting her with a direct directive bald on record: "Hang on β I haven't finished!" However, in a formal context, more indirect language would be expected.
According to Heritage (1984), a speaker's action is context-shaped in that its contribution to an ongoing sequence of actions cannot be understood except by reference to the context. This contextualization of utterances is a major and unavoidable procedure which hearers use to interpret conversational contributions.
π Definition β Context-shaped: A speaker's action whose contribution to an ongoing sequence of actions cannot be understood except by reference to the context. π Rule: The greater the imposition, the more indirect the language is. π Rule: The greater the formality, the more indirect language is. π Example: Large imposition (borrowing Rs.50,000) β "I couldn't borrow Rs.50,000/-, could I, if you don't need it right now?" (indirect, negative politeness). Small imposition (borrowing Rs.10) β "Give me 10 rupees" (bald on record).
Topic β 157 Social Context
The choice of politeness formulation depends on the social distance and the power relation between speakers. When there is social distance, politeness is encoded and there is more indirectness. When there is less social distance, there is less negative politeness and indirectness. Variables that determine social distance include: degree of familiarity and differences of status, roles, age, gender, education, class, occupation, and ethnicity.
The degree of familiarity means that if speakers know each other, they do not need to use politeness strategies; if they use them, it can imply the opposite of politeness. It is those of the lower status, the less dominant role, and so on who use more indirectness and more negative politeness features (such as hedges and mitigation) than those with higher status. Expressions with bald on record are used by people who assume they have power.
π Definition β Social distance: The degree of familiarity between speakers, determined by variables such as status, roles, age, gender, education, class, occupation, and ethnicity. π Definition β Hedges: Linguistic devices used to soften or mitigate the force of an utterance (e.g., 'could', 'possibly', 'if you don't mind'). π Rule: Less social distance β less negative politeness and indirectness. Higher status β more bald on record expressions.
Topic β 158 Cultural Context
According to Tannen (1994), "the use of indirectness can hardly be understood without the cross-cultural perspective." For example, in Cuba, friends should not show any distance at all, and to say 'thank you' for a cup of coffee β "maximizing praise of other" β can cause offense as it appears to put up barriers.
Thomas (1995) mentions that Chinese hosts will choose a guest's menu for them and put the 'choicest pieces' on their plate, to show positive politeness. Similarly, in some Western cultures, refusal demands a specific excuse if speakers are to avoid threatening positive face and "minimize dispraise of other," whereas in other cultures this is not necessary.
π‘ Why this matters: Politeness strategies are not universal; what is polite in one culture may be offensive in another. Understanding cross-cultural variation is essential for effective intercultural communication.
Topic β 159 Applying Politeness
Here is an activity for students. You are supposed to find requesting strategies ranging from the more direct to the more indirect ones: [Can you...?], [Do you think you could...?], [I wonder if you could...?], [Could you...?], [I wonder if you could...?], [Would you...?], [Would it be possible for you to...?]
If the strategies presented do not appear appropriately contextualized, it may be difficult to know how to use them. If a context is provided, the learner has a better idea of how and when to use them. For example, to find out different pragmatic parameters involved behind all these pragmatic choices, answer these questions: (1) Where are the people? (2) What is the relationship between them?
Moreover, meanings emerge from the discoursal context where politeness strategies are used. For example, take the following dialogue:
[FATHER: What sort of time do you call this?] [Boy: I'm sorry.] [FATHER: So you should be! It's 2 a.m.!] [Boy: Oh Dad, do stop nagging. I'm over seventeen. It's up to me what time I come in.] [FATHER: Not while you are living here, it isn't. Anyway, what on earth were you doing until 2 in the morning?] [Boy: We weren't doing anything. We were just talking.] [FATHER: I was worried about you.] [Boy: Honestly, Dad, I really am sorry, but you don't have to wait up for me, you know...]
Students should find what type of politeness strategies are used in this dialogue, and highlight the bald on-record, off record, and positive/negative politeness techniques.
β Key Takeaways
The most critical point is that politeness is a pragmatic phenomenon where meaning depends on function, context, and intention β not just linguistic form. The size of imposition and formality of context determine the level of indirectness in requests. Social variables like status, familiarity, age, and gender influence the choice between bald on-record, positive politeness, and negative politeness strategies. Cultural context is essential because politeness norms vary significantly across cultures, as demonstrated by Cuban and Chinese examples. When analyzing politeness, learners must always consider discoursal context β including participants, setting, and relationship β to correctly identify politeness strategies.
π§ Quick Revision Questions
- What is the difference between politeness and deference? Provide an example of an honorific that signals deference in English.
- According to the lecture, what two situational context factors influence how we make a request, and what is the general rule for each?
- List at least five social variables that determine social distance between speakers.
- Why might saying "thank you" for a cup of coffee cause offense in Cuba, according to Tannen (1994)?
- In the father-son dialogue provided, identify one example of bald on-record language and one example that might involve negative politeness or a face-threatening act.
π Lecture 32 β Conversational Analysis
π Overview: This lecture introduces Conversational Analysis, the study of how people manage ordinary everyday spoken interactions. It examines the structural features of conversation, including turn-taking, pauses, overlaps, backchannels, adjacency pairs, preference structures, and sequences. Understanding these mechanisms is essential for analyzing how meaning is co-constructed in real-time social interaction.
ποΈ Topics Covered
The lecture covers seven main topics: what conversational analysis is and its key features; turn-taking as a locally managed system with cultural variations; pauses, overlaps, and backchannels as conversational signals; adjacency pairs as two-part sequences; preference structure distinguishing preferred and dispreferred responses; and sequences including pre-sequences, insertion sequences, and opening/closing sequences.
π Lecture Summary
Topic β 160 What Is Conversational Analysis?
Conversational analysis examines ordinary everyday spoken discourse to understand how people manage their interactions. It is the study of social interaction embracing both verbal and non-verbal conduct in everyday life. It involves a rigorous investigation of features of a conversation: how it is generated and constructed, how it operates, what its distinguishing features are, and how participants construct their own meanings in the conversational situation.
Conversations are described as multi-layered/multi-leveled, meaning they contain different levels of meaning within a single text. This analysis examines the contents, sequence, evolution, and forms of conversation. We study how speakers decide when to speak during a conversation, how utterances of two speakers are related, and the different functions that conversation serves. Key features of conversational analysis include turn-taking, pauses, overlaps and backchannels, adjacency pairs, preference structures, and sequences expansion.
π‘ Why this matters: Conversational analysis reveals the hidden structure beneath seemingly casual talk, showing that everyday conversation follows systematic rules that participants unconsciously know and follow.
Topic β 161 Turn-Taking
Turn-taking operates in accordance with a local management system that is conventionally known by the members of a social group. This system is essentially a set of conventions for getting turns, keeping them, or giving them away. It is needed most at those points where there is a possible change in who has the turn. Turn-taking can be applied to conversations where speakers cooperate and share the floor equally, as well as to conversations where there is competition, with speakers fighting to keep the floor and preventing others from getting it.
Different cultures have their own preferences regarding how long a speaker should hold the floor and how they indicate they have finished so another speaker can take the floor. For example, Latin Americans have pauses of a fraction of a second and find it socially acceptable to overlap and interrupt, whereas North American Indians expect a two-second pause between turns, and Japanese culture considers it unacceptable to interrupt.
Topic β 162 Pauses, Overlaps and Backchannels
When two people attempt to have a conversation and discover there is no 'flow' or smooth rhythm to their transition, much more is being communicated than is said; there is a sense of distance, an absence of familiarity or ease. Very short pauses (marked with a dash) are simply hesitations, but longer pauses become silence. These silences are not attributable to either speaker because each has completed a turn.
Overlaps are conventionally marked by a double slash (//) at the beginning of the overlapping talk. Typically, the first overlap occurs as both speakers attempt to initiate the talk. In accordance with the local management system, one speaker will stop to allow the other to have the floor. Backchannels are head nods, smiles, and other facial expressions and gestures that indicate the hearer is listening. These types of signals (uh-uh, yeah, mmm) provide feedback to the current speaker that the message is being received; they are known as backchannel signals or simply backchannels.
Topic β 163 Adjacency Pairs
A sequence of two related utterances by two different speakers is known as adjacency pairs. The second utterance is always a response to the first. For instance: [A: You left the light on.], [B: It wasn't me!] β the sequence of complaint-denial forms an adjacency pair.
Types of adjacency pairs include: question-answer, greeting-greeting, invitation or offer-acceptance/non-acceptance, complaint-apology/denial, summons-response, request-acceptance, promise or thanks-acknowledgement, and goodbye-goodbye.
An insertion sequence is one adjacency pair within another. A number of insertion sequences can be infinite, but the limit of human memory does not allow that. Delay in response marks potential unavailability of the immediate expected answer; it represents the distance between what is expected and what is provided.
Topic β 164 Preference Structure
Adjacency pairs represent social actions, and not all social actions are equal when they occur as second parts of some pairs. For example, in response to [a question], possible responses include [a partial answer], [a statement of ignorance], [a denial of the relevance of the question], or [a denial of its presupposition]. For instance: [What does Joe do for a living?] could receive responses like [Do you need to know?], [Oh, this and that.], [I've no idea.], [What's that got to do with it?], or [He does not do anything.]
A first part that contains a request or an offer has an expectation that the second part will be an acceptance; structurally, acceptance is more likely than refusal. This structural likelihood is called preference. Technically, preference is an observed pattern in talk and not a personal wish. The preference structure divides second parts into preferred (structurally expected next act) and dis-preferred (structurally unexpected next act) social acts.
Topic β 165 Sequences
Conversation analysts claim that as speakers are mutually constructing and negotiating their conversation in time, certain sequences emerge β these are stretches of utterances or turns. These can be pre-sequences, insertion sequences, and opening and closing sequences.
Pre-sequences are the opening sequences used to set up some specific potential actions. Conversations are opened in socially recognized ways. In the beginning of a conversation, speakers normally greet each other, e.g., when two office workers meet in the morning: Ali: [Morning, Salman!], Salman: [Hi. How's it going?], Ali: [Oh, can't complain, I guess. Ready for the meeting this afternoon?], Salman: [Well, I don't have much choice!]
Greetings exemplify opening sequences β utterances that ease people into a conversation. They convey the message "I want to talk to you." There are pre-invitations (I've got two tickets for the football match), pre-requests (Are you busy right now?), and pre-announcements (You'll never guess!).
Finally, there are opening structures and closing structures. For example, Iqra, a 34-year-old housewife, greets Laiba, a 15-year-old student, with a formulaic health enquiry: Iqra: [Hi. Laiba.], Laiba: [Hi, Hi, Iqra.], Iqra: [Hi, hi, how are you?], Laiba: [Not bad. I'll be in, in a minute.]
The British and North Americans have a pre-closing sequence rather than just ending with a farewell, e.g., A: [Anyway, I'm gonna have to go.], B: [Yeah, see you.], A: [See you tomorrow.], C: [What time is it?], D: [Oh. I've my lights on.], E: [Half three.], C: [Three.]
β Key Takeaways
Conversational analysis is the systematic study of how people manage ordinary spoken interaction, examining turn-taking, pauses, overlaps, backchannels, adjacency pairs, preference structures, and sequences. Turn-taking operates through a culturally-variable local management system with conventions for getting, keeping, and giving turns. Adjacency pairs are two-part sequences where the second utterance responds to the first, with preference structure distinguishing structurally expected (preferred) from unexpected (dispreferred) responses. Sequences include pre-sequences (which set up future actions), insertion sequences (embedded within other pairs), and opening/closing sequences (which socially ease participants into and out of conversation). Understanding these structures reveals that conversation is not random but follows systematic, analyzable patterns that participants mutually construct and negotiate.
π§ Quick Revision Questions
- What are the six main features of conversational analysis listed in this lecture?
- How does turn-taking differ across Latin American, North American Indian, and Japanese cultures?
- What is the difference between an overlap and a backchannel signal in conversation?
- What is an adjacency pair? Provide three examples of different types of adjacency pairs.
- What does "preference" mean in preference structure, and how does it differ from a personal wish?
π Lecture 33 β CULTURE AND LANGUAGE LEARNING
π Overview: This lecture explores the intersection of culture and pragmatics in language learning, examining how cultural background influences communication across languages. It introduces cross-cultural, intercultural, and interlanguage pragmatics as distinct but related fields that help explain how speakers from different cultural backgrounds interpret and produce meaning in interaction.
ποΈ Topics Covered
The lecture covers cross-cultural pragmatics as the study of linguistic acts across cultural backgrounds, intercultural pragmatics based on socio-cognitive approaches to communication, interlanguage pragmatics focusing on second language learners' pragmatic development, listener behavior in cross-cultural pragmatics including back channel cues, and referential communication as an interactive process for conveying given and new information effectively.
π Lecture Summary
Topic β 166 Cross-cultural Pragmatics
Cross-cultural pragmatics (CCP) is a subfield of pragmatics. Kasper and Blum-Kulka (1993) defined CCP as the study of linguistic acts by language users from different cultural backgrounds. CCP aims at understanding the extent to which non-shared knowledge or, in simple terms, our schemas affect or modify the retrieval of intended meaning.
CCP looks at issues outside classrooms and concentrates on environments where participants are not explicitly learners, but rather full members of the target language community. In pragmatics, "culture" is not Culture with a capital C β that is, the literature, music, and art. Rather, it is culture as a reflection of the values and beliefs about the world, held by the members of a community which forms the substratum of their everyday life.
CCP investigates how human behavior is translated into instances of language in use. Research has shown that a speaker's intended meaning, mediated by linguistic symbols, may be interpreted or misinterpreted in cross-cultural contexts due to each interactant's own norms of interpretation. CCP examines behaviors that are manifest or overt and others that are latent or covert. The values and beliefs are embedded in talk both at the micro and the macro level. Micro features include prosodic cues, turn taking, indirectness, nonverbal cues, etc.
π Definition β Cross-cultural Pragmatics: The study of linguistic acts by language users from different cultural backgrounds, examining how non-shared knowledge and schemas affect the retrieval of intended meaning.
π‘ Why this matters: Cross-cultural misunderstandings often arise not from grammar errors but from differences in cultural assumptions about appropriate communication.
Topic β 167 Intercultural Pragmatics
Intercultural pragmatics is based on the socio-cognitive approach according to which, our mind exists simultaneously both in the head and in the world. Cross-cultural comparative studies of discourse have shown that rules of appropriateness vary across cultures. Thus, for learners to become truly effective communicators in a second language, they need to acquire these rules of appropriateness in addition to linguistic competence.
Modern language courses include elements of 'communicative competence' (Hymes, 1964), providing learners with knowledge about and experience in using the sociocultural rules of the new language. Intercultural pragmatics studies speech acts across various cultures to study how language is manifested in a society and is a medium of differentiation between different cultures. The difference in culture causes a number of intercultural failures such as socio-pragmatic failure and pragma-linguistic failure.
π Definition β Socio-pragmatic failure: Miscommunication arising from different cultural assumptions about social appropriateness. π Definition β Pragma-linguistic failure: Miscommunication arising from inappropriate transfer of linguistic features from one language to another.
π Example: A learner who directly translates a polite request form from their native language into English may produce a phrase that sounds rude or overly formal to native English speakers β this is pragma-linguistic failure.
Topic β 168 Inter-language Pragmatics
Interlanguage pragmatics (ILP) focuses on the pragmatic development of second/foreign language learners as they seek to become proficient, successful users of the target language. ILP may be considered a sub-area of second language acquisition (SLA) and thus is inherently anchored in acquisition contexts. Learners and teachers need to understand interlanguage pragmatics to facilitate our understanding of how people comprehend and communicate meaning beyond what is said in the target language.
The interlanguage is a reduced system at the early stages of development. Interlanguage refers to intermediate, dynamic, and transient linguistic systems that, according to the theory, continue to develop over time as learners move closer to attaining native-like proficiency. Learners progress along a trajectory towards their desired goal, sometimes regressing and sometimes making rapid progress. For instance, in academic speaking skills, the learners may struggle with conversational chat unless they live with a native speaker.
Learner language use is informed by underlying systems that overlap and are constantly being revised by learners, usually out of awareness, as long as there are regular exposure and tutoring. ILP is essentially interested in how L2 learners use their developing abilities in L2 to communicate successfully despite gaps in their knowledge and socio-pragmatics of the L2.
π Definition β Interlanguage: Intermediate, dynamic, and transient linguistic systems that continue to develop over time as learners move toward native-like proficiency.
Topic β 169 Listener Behavior in CCP
One aspect of interactional talk is the relative lack of attention given to the role of the listener in the co-construction of meaning in conversation. Yet it is important not only cross-culturally, but also within a pluricultural society.
Erickson and Schultz (1982) found that African-American students displayed attention and listenership in less overt ways than those expected by white counselors, even when there were no gender differences. Listener cues are also called "back channel cues". Cues are believed to be given by the listener while the speaker, often in the context of telling a story or narrating recent events (Yngve 1970).
Cultural values are attached to cues more explicitly in one culture than another. For example, Japanese culture does transparently discuss the need for the listener to "chime in" in a patterned manner. The cross-cultural mismatches between interlocutors of different languages may arise regarding the location and frequency of listener responses.
π Definition β Back channel cues: Verbal and nonverbal signals given by listeners to indicate attention and understanding while a speaker is talking.
π Example: In Japanese conversation, listeners are expected to produce frequent back-channel responses like "hai" or "un" in a patterned rhythmic way, whereas in some Western contexts, such frequent responses might be interpreted as interrupting.
Topic β 170 Referential Communication
Tarone and Swierzbin (2009: 68) chose to study how learners used language resources to signal given and new information to achieve effective referential communication. They stated: "Effective referential communication is inherently interactive. Speakers must take into account any information the listener already has, and any information the listener needs to know in order to identify the entity being referred to." It is also an aspect of language use that depends on careful decisions by speakers regarding linguistic forms, for example, in Academic Speech Events.
Bardovi-Harlig and Hartford (1990, 1993) focused their analysis on status-preserving strategies in the interactions between the advisors and advisees (international students), particularly in an educational setting where the pragmatic competence of native and non-native speakers of the L2 are tested. They are of the view that highly proficient nonnative speakers may lack the ability to negotiate an advisement interview that is dependent on appropriate speech acts and communication strategies. At issue in the study is impression management, important to graduate students who seek to be well regarded by their advisors.
π Definition β Referential communication: The interactive process by which speakers signal given and new information to enable listeners to identify the entity being referred to.
π‘ Why this matters: Even highly proficient second language speakers may fail to manage institutional interactions (like advisor meetings) because they lack the pragmatic knowledge of appropriate status-preserving strategies.
β Key Takeaways
Cross-cultural, intercultural, and interlanguage pragmatics each examine different aspects of how culture affects language use and learning across cultural boundaries. The key distinction is that cross-cultural pragmatics compares language use across cultures, intercultural pragmatics studies actual interactions between speakers of different cultures, and interlanguage pragmatics focuses on learners' developing pragmatic competence. Listener behavior, particularly back channel cues, varies significantly across cultures and can cause miscommunication even without gender differences. Referential communication requires speakers to carefully manage given and new information while considering the listener's knowledge state. Pragmatic failures can occur at both the socio-pragmatic level (cultural appropriateness) and pragma-linguistic level (linguistic form appropriateness), and even advanced learners may struggle with these aspects.
π§ Quick Revision Questions
-
What is the difference between cross-cultural pragmatics and intercultural pragmatics according to this lecture?
-
What are back channel cues, and why can they cause cross-cultural misunderstandings?
-
What is interlanguage, and how does interlanguage pragmatics relate to second language acquisition?
-
What two types of intercultural failures are identified in the lecture, and how do they differ?
-
According to Bardovi-Harlig and Hartford, why might highly proficient nonnative speakers struggle in advisement interviews?
π Lecture 34 β Computer Semantics and Pragmatics
π Overview: This lecture explores the intersection of computers with semantics and pragmatics, focusing on how computational tools are used to analyze and process language. It covers computer-assisted language learning, computational linguistics, corpus linguistics, machine translation, and the challenges of modeling pragmatics computationally, highlighting both achievements and limitations in the field.
ποΈ Topics Covered
The lecture begins with Computers and Linguistics, discussing CALL and computational linguistics, then moves to Computers, Corpora, and Semantics to explain how corpora reveal patterns of language use. It covers Computers and the Study of Translation, comparing transfer and interlingual models, then examines Computers and Pragmatics, addressing why pragmatics is less explored computationally. Finally, it discusses Computers, Speech Acts, and Conversation, focusing on spoken corpora and their challenges.
π Lecture Summary
Topic β 171 Computers and Linguistics
Computer-Assisted Language Learning (CALL) is defined as the search for and study of applications of the computer in language teaching and learning. CALL covers a broad range of activities, making it difficult to describe as a single idea or simple research agenda; it encompasses issues of materials design, technologies, pedagogical theories, and modes of instruction.
Computational Linguistics (CL), on the other hand, is the study of language from a computational perspective, concerned with computer systems for understanding and generating natural language. The process of dictionary making today is undergoing dramatic change largely owing to advances in computers and the availability of machine-readable collections of texts known as corpora.
Corpus linguistics is the term used for compiling collections of texts and using them to probe language use. A corpus is a representative body of texts (Latin for 'body'), with more recent corpora containing over 100 million words. Corpora are essential for twenty-first-century dictionary making, speech recognition, and artificial intelligence.
Topic β 172 Computers Corpora and Semantics
Computerized corpora are useful to dictionary makers and others in establishing patterns of language that are not apparent from mere introspection. For example, patterns of collocation β which words go together β are much more readily understood with the help of a computerized corpus of natural-language texts. Such patterns help highlight meanings, parts of speech, and words that co-occur with frequency.
π Definition β Collocation: patterns of words that frequently go together in a language, revealed through corpus analysis.
Furthermore, while it may appear that synonymous words can be used in place of one another, corpora can show that it is not common for words to be readily substitutable. For example, little and small, big and large, and fast and quick are generally considered synonyms. But a cursory examination of key word in context (KWIC) concordances for these pairs shows that they are not straightforwardly substitutable. π Example: The synonyms little and small β a KWIC concordance would show they are used in different contexts and are not interchangeable, even though they share similar meanings. Corpora thus help understand meaning and use of words.
Topic β 173 Computers and the Study of Translation
For more than half a century, researchers have been trying to craft devices that can translate between languages, but except in limited ways, that goal has eluded the best attempts. Word-for-word translation does not work because languages differ in their word orders, and metaphors of one language may not translate into the relevant metaphor of another language.
If a machine translation (MT) device were established for even six languages, then 6 Γ 5 = 30 sets of procedures would be needed to translate each language into all of the other five (the transfer translation model). An alternative system is an interlingual translation model, in which the basic semantic elements of each language can be represented abstractly and then encoded into other languages. This would require twelve procedures β one decoding procedure and one encoding procedure for each language. This model is simpler than the transfer translation model.
π Formula: Transfer translation model for n languages β n Γ (n-1) procedures; Interlingual model β 2n procedures. π Example: For 6 languages, transfer model = 30 procedures; interlingual model = 12 procedures.
Unfortunately, we don't know the extent to which sentences can be decomposed into the kinds of abstract semantic representations needed for an interlingual model, especially making a language-neutral intermediate representation. Difficulties concern what one language encodes that another may not encode.
π Definition β Interlingual translation model: a machine translation approach where basic semantic elements are represented abstractly and then encoded into other languages, requiring fewer procedures than the transfer model.
Topic β 174 Computers and Pragmatics
Compared to morphological, lexical, phonological, grammatical, and semantic features of texts, the features of pragmatics have been less well explored in computational linguistics. One reason for this relative neglect is that modeling the world knowledge and discourse knowledge that speakers rely on when producing and understanding texts is even more challenging than creating models of structural aspects of language.
Another reason is that the kinds of linguistic features by which some pragmatic categories are realized are not always expressed in ways that computers can readily track. Speakers base some aspects of expression on their beliefs about what addressees know and are thought to have in the forefront of their minds. π‘ Why this matters: This highlights the fundamental challenge of making computers understand implied meaning, context, and speaker intentions β central to pragmatics.
Topic β 175 Computers, Speech Acts and Conversation
It is necessary to create models of politeness, turn taking, and other conversational phenomena before many applications of computer technology to speech will be mastered. The building of corpora of written language has proceeded more quickly than the compilation of spoken corpora. Machine-readable texts from books, magazines, and newspapers have been widely available, as is the Web. Scanners can effectively transform many older printed materials into machine-readable text.
However, creating electronic representations of transcribed spoken language is quite different. First, it must be captured on audio or video, then transcribed β a challenging and expensive task, partly dependent on recording quality and ambient noise. Still, a substantial part of the British National Corpus is based on speech, where about one hundred volunteers throughout Britain carried recorders.
β Key Takeaways
The lecture fundamentally shows that while computers have transformed linguistic analysis through corpora and computational methods, the handling of meaning and especially pragmatics remains profoundly challenging. Corpora are indispensable for revealing patterns like collocation and synonym use that introspection cannot detect. Machine translation is constrained by language-specific structures and metaphors, with the interlingual model being more efficient but theoretically problematic. Pragmatics is neglected computationally because world knowledge, discourse knowledge, and speaker intent are extremely difficult to model. Finally, spoken corpora are harder to compile than written ones, yet essential for modeling speech acts, politeness, and conversation in computational linguistics.
π§ Quick Revision Questions
- What is the difference between Computer-Assisted Language Learning (CALL) and Computational Linguistics (CL)?
- How do computerized corpora help reveal that synonyms like little and small are not always substitutable?
- Why is the interlingual translation model more efficient than the transfer model for six languages, and what is its main challenge?
- What are two reasons pragmatics has been less explored in computational linguistics compared to other language features?
- Why is compiling a spoken corpus more difficult than compiling a written corpus, and what example of a spoken corpus is mentioned?
π Lecture 35 β Information Structure and Pragmatics
π Overview: This lecture explores how speakers and writers organize information in discourse through foregrounding and backgrounding, examining the concepts of given and new information, topics and contrast, and the distinctions between definite/indefinite and generic/specific expressions. Understanding information structure is crucial for analyzing how meaning is conveyed beyond individual sentences in real communication.
ποΈ Topics Covered
The lecture begins with an introduction to information structure as a discourse-level phenomenon involving highlighting and deemphasizing elements in texts. It then examines the distinction between given information (already in the addressee's mind) and new information (just being introduced). The concepts of topic (what a sentence is about) and comment (what is said about the topic) are explained alongside contrastive noun phrases. Finally, the lecture covers definite and referential expressions, then concludes with the distinction between generic and specific noun phrases.
π Lecture Summary
Topic β 176 Introduction to Information Structure
Speakers and writers are responsible for foregrounding certain elements and backgrounding others in texts, similar to how a painter uses contrasts to highlight some details. In language, this process of highlighting and deemphasizing is called information structure. Unlike syntax and semantics, which focus on individual sentences, information structure requires consideration of discourseβcoherent sequences of sentences rather than isolated ones.
π‘ Why this matters: Understanding information structure helps explain why some sentences feel natural in context while others seem to emphasize the wrong elements at the wrong time.
Topic β 177 Given and New Information
Given information is information currently in the forefront of an addressee's mind. New information is information just being introduced into the discourse.
π Definition β Given Information: Information that can be presumed to be already in the mind of the addressee based on prior discourse or context.
π Definition β New Information: Information being introduced into the discourse for the first time.
π Example: Alina: [Who ate the biscuits?], [Maria: Hassan ate the biscuits.]
- In Maria's answer, "Hassan" represents new information because it is being introduced.
- "the biscuits" is given information because Alina introduced it in the previous turn.
Given information can be realized in condensed formβinstead of "Hassan ate the biscuits," the speaker could say "Hassan did," "Hassan did it," or simply "Hassan."
A piece of information need not be explicitly mentioned to be considered given. Information is sometimes taken as given because of its close association with something introduced into the discourse. When a noun phrase is introduced, all the subparts of the referent can be treated as given information.
π Example: "When Kent returned my car last night, the gas tank was nearly empty and the glove compartment was stuffed with candy wrappers."
- Once "my car" is introduced, its parts (gas tank, glove compartment) become given information.
Topic β 178 Topics and Contrast
The topic of a sentence is its center of attentionβwhat it is about and its point of departure. The comment is the element of a sentence that says something about the topic.
π Definition β Topic: The given information in a sentence about which something is said.
π Definition β Comment: The new information that shows what we say about the topic.
π Example: Question: "What did Hassan do?" Answer: "Hassan ate the biscuits."
- Topic: "Hassan" (given information)
- Comment: "ate the biscuits" (new information)
A noun phrase is said to be contrastive when it occurs in opposition to another noun phrase in the discourse.
π Example: Alan: "Did Matt see the accident?" Beth: "No, Sara did."
- "Sara" in Beth's answer is contrasted with "Matt" in Alan's question.
Contrast is also marked in sentences that express narrowing down a choice from several candidates to one. The noun phrase referring to the chosen candidate is marked contrastively.
Topic β 179 Definite and Referential Expressions
Speakers mark a noun phrase as definite when they assume the addressee can identify its referent. Otherwise, the noun phrase is marked as indefinite.
π Definition β Definite: A noun phrase marked when the speaker assumes the addressee can identify the referent.
π Definition β Indefinite: A noun phrase marked when the speaker cannot assume the addressee can identify the referent.
π Example: Andrea: "Who's at the door?" Bundy: "It's the neighbor."
- "the neighbor" is definite, presupposing Andrea can determine which neighbor.
- If they have several neighbors and Bundy cannot assume identification, the answer would be indefinite: "It's a neighbor."
Proper nouns and pronouns are generally definite. Pronouns like "you" and "we" usually refer to individuals identifiable in the discourse context. Definiteness in English is marked by:
- Choice of articles (definite "the" versus indefinite "a")
- Demonstratives ("this" and "that," both definite)
Indefinite noun phrases are marked by "a" or "an" ("a furnace," "an apartment building") or by the absence of any article ("oil," "fire," "apartment buildings").
Topic β 180 Generic and Specific Expressions
Generic sentences express generalizations about kinds and are important tools for transmitting knowledge.
π Example: Compare the following pairs: 1a. Tigers are striped. (generic β saying something general about all tigers) 1b. Tigers are on the front lawn. (specific β referring to particular tigers)
2a. A tiger is striped. (generic) 2b. A tiger is on the front lawn. (specific β referring to some individual tiger)
3a. The tiger is striped. (generic) 3b. The tiger is on the front lawn. (specific β referring to some unique, salient, or familiar tiger)
A noun phrase may be generic or specific depending on whether it refers to a category or to particular members of a category.
π Example 1: "[It describes the bombardier beetle, which squirts a lethal mixture into the face of its enemy.]"
- "the bombardier beetle" is generic (refers to the set of all bombardier beetles) and definite
- "a lethal mixture" is generic and indefinite
π Example 2: "[The giraffe bent slowly forward and gingerly took a carrot from my palm.]"
- "the giraffe" is specific (refers to a particular animal) and definite
- "a carrot" is specific and indefinite
Thus, the generic/specific contrast differs from the definite/indefinite contrast.
β Key Takeaways
Information structure operates at the discourse level, requiring analysis of coherent sentence sequences rather than isolated sentences. The distinction between given and new information is fundamentalβgiven information is what the addressee already has in mind, while new information is being introduced, and given information can be realized in condensed forms or inferred through association. The topic-comment structure pairs given information (topic) with what is said about it (comment), and noun phrases can be contrastive when opposed to other noun phrases in discourse. Definiteness depends on whether the speaker assumes the addressee can identify the referent, marked by articles and demonstratives, while the generic/specific distinction is about whether a noun phrase refers to a whole category or particular members, operating independently of the definite/indefinite distinction.
π§ Quick Revision Questions
-
What is information structure and why does it require consideration of discourse rather than just individual sentences?
-
In the exchange "Who ate the biscuits?" / "Hassan ate the biscuits," which element is new information and which is given information, and why?
-
What is the difference between a topic and a comment in a sentence, and how do they relate to given and new information?
-
When would a speaker use a definite noun phrase like "the neighbor" versus an indefinite noun phrase like "a neighbor"?
-
Explain the difference between generic and specific noun phrases using "The tiger is striped" versus "The tiger is on the front lawn."