LING/PSYC 370B Issues, Theories, and Methods Lecture Notes PDF

Issues, theories, and methods LING/PSYC 370B January 13th & 16th Gabrielle Manning  2 Week 2 roadmap  Psychological mechanisms  Memory  Theories of language acquisition  Research methods  3 Psychological mechanisms ...

Issues, theories, and methods LING/PSYC 370B January 13th & 16th Gabrielle Manning  2 Week 2 roadmap  Psychological mechanisms  Memory  Theories of language acquisition  Research methods  3 Psychological mechanisms  4 “The black box” Input Output Language acquisition Articulation/production of ideas Recognition  5 Serial & parallel processing  Parallel processing: when a group of processes takes places simultaneously  Serial processing: when a group of processes takes place one at a time Parallel distributed processing (PDP): version of the parallel processing model which states that the mind is “massively parallel”  6 Some examples!  7 Parallel distributed processing How do these examples provide evidence for a parallel processor? Whole word information “Repair” of the obscured letter  8 Top-down & bottom-up processing Discourse/meaning Top-down processing: higher Link meaning of a level processing influences sentence with preceding ones lower levels Syntactic Organize words into Bottom-up processing: phrasal structure incoming information is Lexical processed at lower levels Retrieval of words without influence from higher levels Phonological phonemes syllables symbols  9 Top-down & bottom-up processing Top-down processing! Lexical Retrieval of words Phonological phonemes syllables symbols  10 Top-down processing: speech perception Phoneme restoration: When a phoneme is obscured, listeners fill in the missing sound It was found that the *eel was on the axel It was found that the *eel was on the shoe It was found that the *eel was on the orange It was found that the *eel was on the table (Warren & Warren, 1970)  11 Top-down & bottom-up processing  Top-down processing is not always parallel and bottom-up processing is not always serial! So much talk about processing…today in class Syntactic Organize words into phrasal structure Lexical Retrieval of words  12 Automatic vs. controlled processes  Automatic processes: do not require substantial cognitive resources, unintentional and uncontrolled RED RED  Controlled processes: require substantial cognitive resources, conscious Reading = more automatic Colour naming = requires more control  13 MEMORY  14 Information processing model Maintenance/rehearsal Working Permanent Sensory stores memory memory Input/ stimuli Visual, Current goals, Knowledge, auditory, etc. consciousness, skills, belies, etc. etc.  15 Sensory stores  Take in a variety of sensory input  Store raw information  Separate register for each of the five senses  16 Information processing model Maintenance/rehearsal Working Permanent Sensory stores memory memory Input/ stimuli Visual, Short-term & Knowledge, auditory, etc. working skills, belies, memory etc.  17 Short-term memory The squirrel that the man saved ran up the tree The squirrel that the dog that the man saved chased ran up the tree  Short term memory: part of our memory that stores limited amounts of information for a limited amount of time  7±2 units of information 839-939-3949  18 Working memory Maintenance/rehearsal  Working memory: work with the Working information held in short-term memory memory  Maintenance: hold information in short- Short-term & term memory working memory  Manipulation: work on information  19 Let’s test it!  Memory span test (simple span test)  Numbers  Letters  Words  Backward span task  20 Working memory  Memory span: the number of items that can be reliably recalled in the correct order  Primary and recency effects: higher recall for the first and last few list items  Fragility of storage – distractions increase forgetfulness  21 The Baddeley-Hitch Model (1974)  Model of working memory Visuospatial Phonological sketchpad Central loop  3 main components executive  Central executive  Visuospatial sketchpad  Phonological loop  22 The phonological loop  Phonological store: holds phonological representations Visuospatial Phonological sketchpad Central  Articulatory rehearsal loop executive system: covertly and overtly rehearse materials  23 The visuospatial sketchpad  Maintain and manipulates visuospatial information Visuospatial Phonological sketchpad  Allows us to: Central loop  Form visual images executive  Rotate images in our minds  Convert words into images  Etc…  24 The central executive  Exerts executive control  “Assign” the Visuospatial Phonological sketchpad phonological loop to Central verbally rehearse some loop executive material  25 Permanent memory Maintenance/rehearsal Working Permanent Sensory stores memory memory Input/ stimuli Visual, Current goals, Knowledge, auditory, etc. consciousness, skills, belies, etc. etc.  26 Permanent memory  Permanent memory: encode and store memories that last long periods of time  3 types:  Episodic: memory of an event that happened while one was present  Semantic: memory containing generalized knowledge of the world  Procedural: memory containing information about how to do things  27 Relevance for language processing Working Permanent Sensory stores memory memory Pattern Chunking words Semantic memory recognition for from sentences contains information auditory signals into grammatical on speech sounds and constitutes (e.g., words retrieved during noun, verb pattern recognition phrases, etc.)  28 Theories of language acquisition  29 “The black box” Language Acquisition Device (LAD) Input Output Language acquisition Articulation/production What happens here? What types of of ideas processes take place to acquire language? Recognition  30 Various approaches Biological Generative Social Domain general Language as a linguistic Language cognitive biological process LAD includes development = Cognitive Relation to non- language structure social process mechanisms for language biological knowledge Interaction and language processes Poverty of the social abilities development = the Neural processes stimulus same as other underlying Innate linguistic types of learning language knowledge – not E.g., rule learning, development language specific statistical learning  31 Universal grammar (UG)  If we consider that language is universal and innate, then children need the correct parameters turned on  Language Faculty: an innate language system that all humans are born with that provides the LAD, UG, and the general knowledge speakers possess regarding commonalities between languages  Principles: concepts that are common among all languages  Parameters: possible language-specific settings for principles  32 Grammatical gender  Arbitrary noun classification  “classes of nouns reflected in the behavior of laFEM plageFEM leMASC penuMASC French associated words” (Hockett, 1958, p.231) “the beach” “the tire”  Can be marked/agree with other words outside of the noun (e.g., articles and adjectives)  Agreement: when an article or adjective has the LaFEM telivisiónFEM es rojaFEM Spanish same gender as the corresponding noun “the television is red”  Not necessarily consistent cross-linguistically laFEM mesaFEM derMASC TischMASC Spanish “the table” “the table” German  33 Universal grammar (UG)  If we consider that language is universal and innate, then children need the correct parameters turned on  Language Faculty: an innate language system that all humans are born with that provides the LAD, UG, and the general knowledge speakers possess regarding commonalities between languages (Chomsky, 1995)  Principles: concepts that are common among all languages  Parameters: possible language-specific settings for principles grammatical grammatical gender parameter gender parameter Bonjour! Hello! principles = nouns can carry gender information (e.g., masculine & feminine) laF plageF  34 Various approaches Biological Generative Social Domain general Language as a linguistic Language cognitive biological process LAD includes development = Cognitive Relation to non- language structure social process mechanisms for language biological knowledge Interaction and language processes Poverty of the social abilities development = the Neural processes stimulus same as other underlying Innate linguistic types of learning language knowledge – not E.g., rule learning, development language specific statistical learning  35 Statistical learning  8-month-olds can segment streams of speech Familiarization bidaku golabu dutaba bidaku vs. bidakupa? padoti  Transitional probability: the probability that a certain Test sequence will occur based on the occurrence of another sequence bidaku dakugo golabu buduta 0.33 0.33 words from parts of two familiarization words bidaku padoti golabu padoti 1 1 1 1  36 Rule learning Training  Infants can learn rules ABA ABB ga ti ga ga ti ti  Preference for the opposite pattern than li na li li na na what they were exposed to initially  Training ABA = preference for ABB  Training ABB = preference for ABA Test ABA ABB wo fe wo wo fe fe  37 Nature vs. Nurture “Language learning is not really something that the child does; it is something that happens to a child placed in an appropriate environment, much as the child’s body grows and matures in a predetermined way when Nativism provided with appropriate nutrition Empiricism and environmental stimulation.” - Preexisting Blank slate Chomsky, 1973 structure to Knowledge, interpret reason, language experience come from experience  38 Nature vs. Nurture Nativism Interactionist Empiricism Preexisting Blank slate Experience + structure to Knowledge, innate interpret reason, language characteristics experience come from Language input experience is important for language development  39 Research methods  40 Research methods  Methods can focus on language compression, production, or both! Longitudinal studies Observational Developmental language patterns Cross-sectional studies Relationship between language Correlation studies development/non-linguistic factors Elicit a specific phenomenon or Experimental studies investigate comprehension abilities  41 Online vs. offline Offline techniques: tap into information about Online techniques: tasks that allow competence you tap into processing (“online”) After processing routines information Judgements and intuitions Real-times processing Questions about language structure Location of processes Metalinguistic knowledge: conscious use of Questions related to performance language  42 Key terms  Habituation: decrease in response to a stimulus due to repeated exposure  Familiarization: exposure to stimuli that will be in the test phase; practice with the task  Test: a single presentation of a stimulus (a single measurement)  Block: a group of tests  (Pseudo) randomization: make the sequence of stimuli random (or pseudo-random) according to the critical experimental conditions  Order: the different randomizations  List: the inventory of stimuli presented to the participant  43 Example! PARTICIPANT 1 PARTICIPANT 2 PARTICIPANT 3 PARTICIPANT 4 TRIAL LUMIÈRE DOG DOG LUMIÈRE CHIEN LIGHT LIGHT CHIEN OREILLE BELL BELL OREILLE CLOCHE EAR EAR CLOCHE FEUILLE PENCIL PENCIL FEUILLE CRAYON LEAF LEAF CRAYON PAYS BOOK BOOK PAYS LIVRE COUNTRY COUNTRY LIVRE BLOCKS ORDER LISTS LISTS  44 Comprehension tasks  Offline → norming studies (standardized assessments)  Online → implicit comprehension/processing  Standardized assessments:  MacArthur-Bates Communicative Development Inventories (MB-CDI)  Peabody Picture Vocabulary Test (PPVT)  Mean length of utterance (MLU)  45 MacArthur-Bates CDI  Parental report  Vocabulary, comprehension, production, gestures, grammar  Strong reliability and validity  46 PPVT  Measure of receptive vocabulary puppy  47 MLU  Average length of utterances in a speech sample  MLU increases as age increases “Mommy’s going downstairs” =5 1 1 1 1 1 + “The truck went vroom” =4 1 1 1 1 9 total # morphemes 9 MLU = = 4.5 total # utterances 2  48 MLU exceptions  Unintentional repetitions “he he is there" =3  Compound words “doghouse” =1  Non-linguistic vocalizations “choo, choo” =0  Proper names “Mickey Mouse” =1  Fillers “umm” =0  Contractions “don’t” =1  Imitated utterances Adult: “Do you want the ball? Child: “Want ball” =0  49 What’s the MLU? “ This one better, but this not better” “Where is mama?” “Don’t touch” “Downstairs” “I go get a pencil and write” “I want umm drink please” “I, I want the toy”  50 High amplitude sucking  Infants suck on a pacifier to hear a sound  Sucking harder = sound Cool, this is Meh… new! Boring… Interesting! sucking sucking to have to hear sound sound stop  51 High amplitude sucking Baseline Habituation Test amplitude test da ba interested/different uninterested/ the same  52 Head-turn preference paradigm  Used to look at infants’ speech or visual preferences and/or learning  Measurement = the amount of time infants orient their attention to a particular sound or object Familiarization Test Your bike always stays in the garage bike The girl rode her big bike dog The bell on the bike was really loud (Jusczyk & Aslin, 1995)  53 Eye-tracking  Eye-tracker: a device that takes into account the position and movement of a participant’s eye gaze  Online measure  54 Visual world paradigm “click on the candy”  How we interact with language and the visual world  Realistic or semi-realistic scenes, a series of objects  Measurements:  Proportion of looking: fixation on a target  Time-course analyses: looking to images as time unfolds  55 Visual world paradigm “The boy will move the cake” “The boy will eat the cake”  How we interact with language and the visual world  Realistic or semi-realistic scenes, a series of objects  Measurements:  Proportion of looking: fixation on a target  Time-course analyses: looking to images as time unfolds  56 Electroencephalography (EEG) Electroencephalography (EEG) Online measure Electrical brain activity captured at the scalp via small electrodes Captures the macroscopic surface activity Can see when groups of neurons fire  57 Event-related potentials (ERPs) Event-Related Potentials (ERPs)  Deflections in the continuous EEG signal in relation to an event  Language components are very, very small!  Measured by peak amplitude – the maximum height of a peak in microvolts  58 ERPs  Electrodes filled with conductive gel  59 ERP components Mismatch N400 P600 Negativity (MMN) The pizza is too hot to drink The broker persuaded to sell ba ba ba ba ga ba ba the stock was tall Spontaneous speech 60  Investigate natural speech production – little to no prompting Language environment analysis (LENA) How much speech do we need to record for it to be representative of natural speech? Minimum = 50 utterances No maximum – research dependent  61 CHILDES Child Data Language Exchange System (CHILDES)  Corpus of spontaneous speech data  Transcripts, audio, video  Cross-linguistic information  Over 200 speech samples from children of various language backgrounds  62 Elicited production tasks  Elicited production tasks: examiners create a context using verbal and non-verbal prompting  Repetition/elicited imitation tasks: participants are asked to repeat what they hear  Used to target specific structures/utterances What What is happening? is the giraffe doing? “The giraffe “There is someone who is is giving someone giving a bath” the giraffe a bath”  63 ManyBabies  100+ researchers  70+ research labs https://manybabies.github.io/  64 So many methods! There are plenty of different ways to tap to investigate language development! It is up to the researcher to decide which task/method suits their research question