Language Notes PDF
Document Details
Uploaded by PleasantFern3720
Tilburg University
Tags
Summary
These notes cover the topic of language, including its structure, the role of grammar, and various theories and methods of studying language. The notes discuss the complexity of human language and its connection to concepts and thought.
Full Transcript
10/9/2024 Chapter 7 Language Language: a system for communicating with others using signals that convey meaning and are combined according to the rules of a grammar. But what is ‘grammar’? -…The systematic description of the feat...
10/9/2024 Chapter 7 Language Language: a system for communicating with others using signals that convey meaning and are combined according to the rules of a grammar. But what is ‘grammar’? -…The systematic description of the features of a language like phonology (sound), morphology (system of word formation), syntax (patterns of word arrangement), and semantics (meaning)? → If so, all spoken languages have grammar. - …A set of rules that specify how the units of language can be combined to produce meaningful messages? → If so, all languages have grammar 283 283 10/9/2024 Chapter 7 Language Human language - is extremely complex: the range of ideas and concepts we can express is far wider than in any other (known) species - can refer to abstract intangible concepts that do not really exist in the physical world “Metaphysics refers to the studies of what cannot be reached through objective studies of material reality” - is used to name, categorize and describe things to ourselves when we think, which influences the way in which thought and knowledge are organized in the brain “I see, well, then tomorrow I’ll travel 42.8 meters to the North-West where those beautiful Lettuces are growing… I should make it just in time for lunch” 284 284 10/9/2024 Chapter 7 Language The study of language and its rules is called Linguistics. It attempts to describe languages in a compact set of rules - For example: the regular plural rule in English = singular + ‘s’ → hat(s), cat(s), sock(s), …. The study of human use of language is called Psycholinguistics. It attempts to describe how we acquire, use, generate and comprehend language (and how this is organized in the brain) - For example: Is the regular plural rule engrained in our brain, or do we memorize the plural form of each word separately? → probably rule based (as, for example, indicated by the wug-test) 285 285 10/9/2024 Chapter 7 Language Language is a hierarchical system with the phoneme at its base. A phoneme is the smallest unit of sound that differentiates between words (‘pat’, ‘cat’). - Phonemes are sounds, not letters (‘pat’, ‘bet’ → different letters, same sound!) - The number of phonemes used across languages varies between 12 and 85 (English and Dutch have about 40) An acoustic distinction does not have to be a phonemic distinction - In Xhosa, different ‘clicks’ are used. The acoustic difference is clear, but these phonemes carry no (or a different) meaning for people who do not use these sounds in their language. - In Japanese, the /r/-/l/ difference is not a phonemic contrast. Japanese has a blended sound between /r/, /l/ and /d/. Japanese listeners can however, often hear the difference between English /r/ and /l/ (depending on age of English acquisition and training) but may have trouble producing it (and English speakers mostly don’t get the Japanese consonant right either). - Spanish 6-month-old infants can match a /ba/ or /va/ sound with a face that articulates that syllable, but can no longer do so at 11 months of age because the /b/-/v/ contrast does not exist in Spanish (Pons et al., 2009). 286 286 10/9/2024 Chapter 7 Language Phonemes form higher order structures such as syllables, morphemes, words, phrases and sentences Phonemes represent sounds, not letters. The International Phonetic Alphabet uses symbols to refer to those sounds (IPA: [saʊnds]) 287 287 10/9/2024 Chapter 7 Language Phonological rules indicate how phonemes can be combined to produce speech sounds (in English, ‘spr’ can occur at the beginning of a word/syllable [spring, spray] but not at the end) and how they are ordered into syllables (a syllable must have a vowel) and morphemes. Morphemes are the smallest meaningful units of a language. - ‘dogs’ contains 2 morphemes: ‘dog’ + plural ‘s’ - free morphemes can stand on their own as a word (like ‘capital’ in ‘capitalism’) - bound morphemes cannot stand on their own as a word (e.g., suffixes like ‘-ness’, ‘-ly’, and prefixes like ‘un-’, ‘re-’) Morphological rules indicate how morphemes can be combined to form words 288 288 10/9/2024 Chapter 7 Language Words are stored in the mental lexicon that has about 40,000 entries (for adults) and informs us about what words are ‘legal’ in a language. This helps us to perceive speech correctly: - When the initial phoneme in a spoken word is ambiguous, say, somewhere in between a ‘g’ and a ‘k’ (denoted as ‘?’) and that sound is embedded in ‘?ift’ or ‘?iss’) we perceive it as ‘g’ when followed by ‘ift’ and as ‘k’ when followed by ‘iss’ because ‘gift’ and kiss’ are words, but ‘kift’ and ‘giss’ are not (so-called Ganong-effect, Ganong, 1980) The y-axes indicate the proportion of ‘k’- responses on sounds from a continuum between ‘g’ (low VOT) and ‘k’ (high VOT). The proportion of ‘k’-responses increases with VOT, but are also higher when ‘k’ formed a word and ‘g’ a non-word (orange line) - Figure from Giovannone and Theodore (2021) 289 289 10/9/2024 Chapter 7 Language If you are multilingual, you (probably) have language-specific lexicons for each language that you know. If a word is activated in the input lexicon, it activates semantics (meaning of a word), syntax, orthography (writing), grammatical gender (depending on language), etc. Clear evidence for semantic processes is obtained in semantic priming studies: - Consider an experiment in which the word ‘dog’ is either followed by the word ‘bone’ or ‘book’. Typically, participants will respond faster and more accurately on ‘bone’ than on ‘book’ → The semantic network that is activated by ‘dog’ includes ‘bone’ but not ‘book’, which facilitates word recognition/processing of ‘bone’. - The word ‘arrival’ has the word ‘rival’ embedded in it, and ‘arrival’ can therefore prime a target word related to ‘rival’, such as ‘enemy’ (Zhang & Samuel, 2015). 290 290 10/9/2024 Chapter 7 Language Syntax (Syntactical rules) specify how the units of language can be combined to form phrases and sentences (see page 283 in your book for example figure). This way, we can make a distinction between poorly formed and well formed sentences, even if they are meaningless. Syntax versus semantics: - My frozen soup is on the stove - syntactically and semantically correct - Frozen soup my on the stove is - syntactically incorrect but meaning can be derived “If seen me before you have, you know that meaning from syntactically strange sentences derived can be” - My frozen cup sleeps in the stapler – syntactically correct but meaningless - Frozen my the sleeps cup stapler in – syntactically incorrect and meaningless 291 291 10/9/2024 Chapter 7 Language In spoken language, information can be derived from prosodic cues like - intonation (pith rises in question) - volume (intensity) - speaking rate (fast vs. slow) Prosody aids in syntactic grouping (“The dog who I hate is gone”) and is important for expressing emotions in the voice. For example, fearful expressions are produced faster and with higher pitch than happy expressions. Knowing this allows researchers to create emotional prosodic cues that range from one emotion to the other (e.g., Baart & Vroomen 2018). Every language has its own prosodic rules. For example, Dutch people accentuate words in a sentence with a pitch rise of ~6 semitones. In Finnish this is lower, and for English, it is larger. 292 292 10/9/2024 Chapter 7 Language Research on the perception of speech sounds has shown that consonants and vowels are perceived categorically (akin to the perception of colors). Task: green or yellow? Task: /ba/ or /da/? % ‘yellow’ responses % ‘ba’ responses The point in the curve where people are guessing reflects the phoneme boundary /ba/ /da/ Auditory continuum - The adjacent stimuli on the continuum are physically equally different from each other - Yet, the S-shaped identification curve indicates that we perceive them as belonging to 2 different categories 293 293 10/9/2024 Chapter 7 Language Furthermore, during discrimination (‘Were 2 stimuli the same or different?’), two stimuli are more easily distinguished if they were identified as belonging to a different category during identification. The idea is that we perceive (or memorize) the category to which a sound belongs (/ba/ or /da/), but not the subtle differences in-between. 294 294 10/9/2024 Chapter 7 Language Research has shown that babies have categorical perception from a very early age ( ‘all’) - verbal short-term memory tasks (“remember 7,3,8,2,4,9,1 in correct order”) - Less well-defined phonetic categories Baart et al. (2012): Identifying sounds on an /aba/ - /ada/ continuum yielded shallower slopes for dyslexics than typical readers (see also e.g. Bogliotti et al., 2008, de Gelder and Vroomen, 1998, Godfrey et al., 1981, Vandermosten et al., 2010, Werker and Tees, 1987) 314 314 10/9/2024 Chapter 7 Language Reading and writing One of the basic problems is that people with DD have trouble connecting the sounds that make up words with the letters that represent those sounds https://www.youtube.co m/watch?v= QrF6m1mRs CQ&t=305s 315 315 10/9/2024 Chapter 7 Language and Thought According to Whorf (1956) “Language is not just a reporting device for our experience, but a defining framework for it”. Language thus determines our experience (linguistic determinism hypothesis). He based himself on the (false) notion that Inuit have more words for ‘snow’ than the English do, and therefore must experience ‘snow’ differently. Against this idea, was the finding that the Dani (Ndani) people from Papua (Indonesia) have 2 color categories (mili for cool/dark shades and mola for warm/light colours) but can nonetheless learn different shades of color as well as people with many more color labels in their language (Rosch, 1973). 316 316 10/9/2024 Chapter 7 Language and Thought However, more recent work with the Himba (a cattle-herding tribe from Namibia) shows that language indeed may influence the way we think and perceive (linguistic relativity hypothesis) English has 11 color names, whereas the Himba have 5. Young English/Himba kids tend to confuse colors that are similar, but as they grow older, performance becomes in-line with the color-names in their native language (Roberson et al., 2004). More support for this idea was found by Boroditsky (2001) Western people describe time as ‘back’ and ‘forward’ → a horizontal time-line Mandarin Chinese people often describe time with ‘up/above’ and ‘down’→ a vertical time-line Time judgements (“March comes earlier than April”) were faster for Western participants after horizontal primes ( ) , whereas Chinese participants were faster after a vertical prime ( ). Although the deterministic view by Whorf seems to be a bridge too far, language does modulate perception when assessed in performance tasks where we need to categorize items. 317 317 10/9/2024 Chapter 7 Language and Thought Concepts and Categories Words refer to concepts. Concepts are mental representations through which we organize information into categories. Some concepts are unambiguous and clear (circle, square), others are less clear (‘cup’ versus a ‘mug’). The question is how we learn and represent concepts. 318 318 10/9/2024 Chapter 7 Language and Thought Concepts and Categories Classical view: Concepts are sets of rules that specify necessary and sufficient conditions for category membership - Necessary: must be true in order to belong to the category (if an object is not a type of fruit, it cannot be an apple) - Sufficient: if true, the object must belong to the category (a ‘Granny Smith’ must be an apple) This is rather artificial (just try to come up with rules that define ‘birds’ that include all birds and exclude all non-birds). 319 319 10/9/2024 Chapter 7 Language and Thought Concepts and Categories Family resemblance theory: members of a category have certain characteristic features, but not every member needs to possess all of these features, and some features are never shared (Wittgenstein, 1953). → What is a “game”? Picture by Allan Sikk: https://www.researchgate.net/figure/Classical-and-family- resemblance-approach-to-concepts_fig2_309573076 320 320 10/9/2024 Chapter 7 Language and Thought Concepts and Categories Prototype theory (Rosch, 1975): categorization is organized around the properties of the most typical member of the category. - Membership is ‘graded’; some objects fit the prototype better than others. 321 321 10/9/2024 Chapter 7 Language and Thought Concepts and Categories Exemplar theory (Medin & Schaffer, 1978): an object is compared with stored memories of all category members (exemplars) we have encountered. Prototype theory Exemplar theory 322 322 10/9/2024 Chapter 7 Language and Thought Concepts and Categories Natural concepts are organized into a hierarchy. The ‘basic’ level is most informative and used most often in conversations “Memory and Learning, from brain to behavior", Mark A. Gluck Semantic network models are often envisioned as following hierarchic structures. When hearing/reading a word, activation spreads throughout the network: - ‘bread’ = strong prime for ‘baker’ - ‘bread’ = weaker prime for ‘knife’ - ‘bread’ = no prime for ‘chalk’ 323 323 10/9/2024 Chapter 7 Language and Thought Concepts and Categories Brain injuries can lead to category-specific deficits: objects that belong to a particular category are not recognized (e.g., human-made objects), whereas items from other categories are recognized just fine (e.g., food, living things). With fMRI, one can further demonstrate category- specific processing in the temporal lobe when subjects silently name pictures of animals versus tools. It thus appears that the brain is pre-wired to group objects into broad- based categories such as living versus non-living things. 324 324 10/9/2024 Chapter 7 Language and Thought Decision making Rational choice theory: We make decisions by determining the value of an outcome and multiply that with the likelihood of occurring. What would you chose? All choices are A) 1,000,000 Euro at 0.0001% likelihood equally rational B) 50 Euro at 2% likelihood (value * likelihood is C) 1 Euro at 100% likelihood the same in all cases) Behavior however, is often not rational. The highly influential prospect theory by Kahneman and Tversky assumes that losses are perceived ‘heavier’ than equivalent gains. - 100% chance to gain $500 or 50% chance to gain $1100 → we mostly pick $500 scenario (risk averse) - 100% chance to lose $500 or 50% chance to lose $1100 → we mostly pick $1100 scenario (risk seeking) 325 325 10/9/2024 Chapter 7 Language and Thought Decision making People are good at judging frequencies, but not very good at judging proportions because we often do not take into account the base-rate (or prior). - Chance of being involved in e-bike accident has increased with 12% over the past 5 years - Chance of being involved in regular bike accident has increased with 10% over the past five years. - Buy a regular bike instead of an e-bike? E-bikes Regular bikes - 5 years ago, 1% of bike owners had an - 5 years ago, 35% of bike-owners had an accident, now it’s 13% accident, now it’s 45% - There are currently 700.000 e-bike - There are currently 700.000 regular bike owners (so 91,000 will have an accident) owners (so 315,000 will have an accident) 326 326 10/9/2024 Chapter 7 Language and Thought Decision making Availability bias: Information that is more readily available, has a large impact on our decision. It is judged as having occurred more frequently and as being more representative. For example: - You do not let your child play outside after seeing news coverage of several high-profile child abductions over the last month. - You refuse to swim in the ocean after seeing several TV shows about vicious shark attacks. - It’s been unusually cold the past week, and you start to wonder if climate change is real. Confirmation tendency: More value is attributed to information that supports a presumption than information that disproves it - E.g., people believe that ‘German cars are reliable’, although they spend more time in the garage than Japanese cars. 327 327 10/9/2024 Chapter 7 Language and Thought Decision making The probability of two events occurring together (in conjunction) is always less than, or equal to, the probability of either one occurring alone - If bank teller and feminist both have a proportion of 1 → probability of conjunction = 1*1 = 1 - If bank teller has a probability of.06 (6%) and feminist has a probability of.90 (90%) → probability of conjunction =.06 *.90 =.054 Conjunction fallacy: believing that two events are more likely to occur together than in isolation. Related to Representativeness heuristic: making a probability judgement by comparing something or someone to a prototype. 328 328 10/9/2024 Chapter 7 Language and Thought Decision making Another statistical misconception: - Gambler’s fallacy: 329 329 10/9/2024 Chapter 7 Language and Thought Decision making Framing effects: the way in which situation is phrased shapes the decision. - 75% chance on success versus 25% chance of failure. Sunk-cost fallacy: people make decisions based on previous investments (re-framing the problem may lead to another decision). Your 2-week camping holiday in France costs you 1195 euro’s that you’ve paid in full in advance. The day before you leave, the weather in France is terrible and a lot of heavy rain is forecasted for the next 18 days. Go on a holiday Stay home - I have invested a lot of money so I need to - The money is already gone, but why go and just make the best of it would I go and be miserable if I can also comfortably stay at home? 330 330 10/9/2024 Chapter 7 Language and Thought Decision making Distractors influence choice - Option 2 is more attractive in menu B Menu A Menu B Option 1 = 20 Euro Option 1 = 20 Euro Option 2 = 35 euro Option 2 = 35 euro Option 3 = 65 Euro 331 331 10/9/2024 Chapter 7 Language and Thought Risky decisions: the Iowa Gambling task Participants choose cards from four different decks. Each card represents an amount of money that can be gained or lost. Two of the decks have low monetary rewards but lower losses. The other two decks have high rewards but even higher losses. Most participants begin the task by picking cards from the disadvantageous deck, but after a few trials, they should see the pattern and begin choosing from the low gain/low loss decks only, thus avoiding losses. Patients with lesions in the ventromedial prefrontal cortex (vmPFC) do not learn this. They pick from both decks equally often. 332 332 10/9/2024 Chapter 7 Language and Thought Problem solving Not only humans have problems that can be solved with insight or thinking. 333 333 10/9/2024 Chapter 7 Language and Thought Problem solving A heuristic is an efficient ‘rule of thumb’ (a ‘shortcut’) that is practical to use, but may not necessarily lead to the optimal answer - broadly applicable to everyday problems - produce generally sound solutions - simplify complicated mental tasks Kahneman: We have 2 systems for solving problems (making decisions). System 1 relies on general observations and quick evaluative techniques (heuristics), system 2 (slow thinking) requires conscious, continuous attention to carefully assess the details of a given problem and logically https://www.simplypsychology.org/wh at-is-a-heuristic.html reach a solution. 334 334 10/9/2024 Chapter 7 Language and Thought Problem solving An algorithm is a procedure that is unique to every situation and always leads to the correct solution (e.g., finding shortest route by checking ALL possible routes). We rely on heuristics instead of algorithms when - the problem space is too large - the time to make a decision is limited - the decision is unimportant - we have access to very little information - an appropriate heuristic comes to mind (by ‘accident’) 335 335 10/9/2024 Chapter 7 Language and Thought Problem solving Analogical problem solving: finding a problem (source) that is similar to the problem you need to solve (target) and applying the solution of the source onto the target. Means-end-analysis (Duncker, 1945): Generating sub-goals to reach desired goal - Goal: get car to start - Subgoals: check gas, check battery, etc. https://www.educba.com/academy/wp- content/uploads/2020/01/Means-Ends-Analysis.jpg 336 336 10/9/2024 Chapter 7 Language and Thought Problem solving As noted, framing affects our decisions, but also our problem solving strategies. Connect all dots with a maximum of 4 straight lines without lifting pencil from paper You have a small candle, some drawing pins, and a box of matches. How can you mount a lit candle on the wall? Functional fixedness is our tendency to perceive an object’s function as fixed. 337 337 10/9/2024 Chapter 7 Language and Thought Problem solving Syllogistic reasoning (related to deductive reasoning): determine whether a conclusion follows from 2 statements that are assumed to be true. People struggle with this as they are often guided by the plausibility of the situation. 1. All children are a nuisance 1. No kids in this school are named Jezebel 2. Jezebel is a child 2. Some kids are named Jezebel Conclusion: Jezebel is a nuisance Conclusion: Some kids are not in this school This is valid reasoning This is valid reasoning 1. All children are a nuisance 1. No kids in this school are named Jezebel 2. Jezebel is a nuisance 2. Some kids are not in this school Conclusion: Jezebel is a child Conclusion: Some kids are named Jezebel This is plausible but invalid reasoning This is plausible but invalid reasoning 338 338