PS: Introduction to Psycholinguistics Winter Term 2005/06 Instructor: Daniel Wiechmann Office hours: Mon 2-3 pm Email: daniel.wiechmann@uni-jena.de Phone: 03641-944534 Web: www.daniel-wiechmann.net Session 3: Visual processing Experience Object Recognition Perceptual organization Sensory input Buttom-up processing - Visual processing originates from sensory input Session 3: Visual processing Prior Experience Object Recognition Top-down processing Perceptual organization Sensory input Session 3: (visual) word recognition Word level Letter level Feature level CAT Assumption: word recognition sequential buttom-up-process Session 3: (visual) word recognition Session 3: (visual) word recognition Word level Letter level top-down effects (word superiority effect) Feature level buttom-up effects Stimulus: COAT Session 3: (visual) word recognition Session 3: (visual) word recognition Summary - word recognition is a combination of buttom-up sensory information and top-down knowledge - word recognition is bi-directional (not sequential) and graded (not discrete) - interactive activation model violates the sequential and discreteness assumptions of a strict information processing model Session 3: (visual) word recognition/ methods Methods to explore visual recognition Brain imaging Examining eye movements Word identification tasks Categorisation times Tachistoscopic identification Session 3: (visual) word recognition/ methods word identification techniques Naming task Subjects name visually presented a words Naming latency is measured (RT ~ 500ms from onset of stimulus) Lexical decision task Subjects decide whether string/sequence is a word or not RT and error rate is measured Session 3: (visual) word recognition/ methods Eye movement in reading e.g. Limbus tracking Infra red beam is bounced off the eyeball and tracks the the boundary between the iris and the white of the eye (limbus) Session 3: (visual) word recognition/observations Reading involves rapid ‘jumps’ called saccades (25 - 60 ms in duration); length is about eight letters 10% of all saccades move backwards Average fixation times range between 200-250 ms Information retrieval takes place in that interval Average span: 15 to the right, 3-4 left (for left to right processing) Session 3: (visual) word recognition/observations Session 3: Towards a model of reading a simple model model Readers fixate on a word until they have processed it sufficiently Then eyes move to the next word Session 3: Towards a model of reading But... Only 80% content words are fixated Only 20% of function words are fixated Rare words are fixated longer than common words Words that are more predictable in sentence context are fixated for less time Words that are not fixated tend to be common, short, or predictable Fixation time of a word is longer when it is preceded by a rare word (spillover effect) Session 3: Towards a model of reading Problems for the simple model: It is hard to see how readers could skip words It takes about 150-200ms to execute an eyemovement program -> readers would waste time waiting for their eyes to move Session 3: Towards a model of reading Advantages eye-movement recording: It provides a detailed on-line record of attentionrelation processes Unobstrusive Disadvantage Hard to to be sure exactly what processing occurs during each fixation Session 3: Towards a model of reading E-Z reader model (Reichle 1998) Readers check frequency (F) of fixated word Completion of F-check is the signal to initiate eye-movement program Session 3: Towards a model of reading E-Z reader model (Reichle 1998) Readers also engage in lexical access (identify orthographic and/or phonological pattern so that semantic information can be retrieved) Completion of lexical access is signal for shift of attention to the next word Session 3: Towards a model of reading E-Z reader model (Reichle 1998) cont.: F-check and lexical access are faster for common words (due to organization of mental lexicon) F-check and lexical access are completed faster for predictable words Session 3: Towards a model of reading: E-Z reader model Time between successive eyemovements in ms 350 300 eye-movement executed completion of frequency check completion of lexical access 250 200 150 100 50 0 1 2 3 4 5 6 7 8 9 10 11 Frequency Effects of word frequency on eye-movements Session 3: Towards a model of reading: E-Z reader model Parafoveal processing Readers spend time between completion of lexical access to a word and next saccade in parafoveal precessing of the next word (this way the model can explain spillover effect) Session 3: (visual) word recognition/observations fovea ~ most sensitive part of the visual field (2 degrees either side of fixation point parafovea (extending 5 degrees) periphery Session 3: Automatic processing Word recognition is fairly automatic Reading is mandatory (cf. Stroop effect) How many mechanisms are involved? Automatic processes: (fast, parallel, not prone to interference from other tasks, cannot be prevented, facilatatory) Attentional (controlled) processes: slow, serial, error prone, uses up working memory (WM), often availble to consciousness, can involve inhibition) Session 3: Priming Priming Involves the presentation of an item A (prime) before reaction to item B (target) is measured stimulus-onset asynchrony (SOA) facilitation vs. Inhibition Form-based priming Semantic priming Session 3: Priming Context effects Semantic (associative) priming Lexical decision task Decision time for target is shorter when prime is semantically related (e.g. DOCTOR - NURSE) Session 3: Priming Priming from sentential context “It is important to brush your teeth every single ___!”