Infant Development VoE Paradigm PDF

PY40008 & PY50408 INFANT DEVELOPMENT The VoE Paradigm  Main Topics Today  Effects of habituation/familiarisation  VoE uses a minimalist perception paradigm  Smart infants and dumb toddlers  Rich interpretation  Improving VoE methods ...

PY40008 & PY50408 INFANT DEVELOPMENT The VoE Paradigm  Main Topics Today  Effects of habituation/familiarisation  VoE uses a minimalist perception paradigm  Smart infants and dumb toddlers  Rich interpretation  Improving VoE methods  Violation-of-Expectation Paradigm – Physical Knowledge  Observation of visual events means it is possible to study infants too young to manipulate objects  May involve an initial event to familiarise infant with the procedure and materials  Initial event also provides key information  In some studies infants are habituated to the initial event Baillargeon, 1993, Chapter  Violation-of-Expectation Paradigm – Physical Knowledge  Test events  Physically possible  Physically impossible  Longer looking at the impossible event indicates  Understanding of the physical principle  Physical reasoning  Mental representation if event was hidden from view (object permanence) Baillargeon, 1993, Chapter  Problems With The VoE Paradigm  We know that looking times can be influenced by many factors other than noticing a violation of a physical law  A priori preferences  Complexity  Movement  Initial familiarisation  Can induce familiarity or novelty preferences  The Drawbridge Study Habituation 1800 Infants looked Impossible longer at the 1800 impossible event Possible 1200 Baillargeon et al, 1985, Cognition, 20, 191-208  The Drawbridge Study Habituation Possible 1200 Event Impossible 1800 Event Baillargeon et al, 1985, Cognition, 20, 191-208  The Drawbridge Study - Object Permanence & Solidity at 5 Months?  5 month infants looked longer at the impossible event  Control – block placed beside drawbridge  No difference in looking times in the control experiment  Seems convincing? Baillargeon et al, 1985, Cognition, 20, 191-208 Baillargeon’s Interpretation: Infants Had A Representation Of The Hidden Block  Representation included information about the block’s:  Position  Height  Physical properties (solid)  Infants knew that solid objects cannot occupy the same space at the same time  This physical law applied to the represented object  Infants engaged in physical reasoning – they expected the drawbridge to stop when it made contact with the hidden block Baillargeon et al, 1985, Cognition, 20, 191-208  Perceptual Differences In The Test Events  Greater change in the impossible 180 degree event may be more interesting  Possible: object, screen  Impossible: object, screen, no object  Noticing the change from object to no object does not mean infants assumed the hidden object continued to exist  For example, you would notice if you saw no reflection in a mirror, but that does not mean you think your reflection is permanent Meltzoff & Moore, 1998, Inf Beh Dev, 21, 201-235  Huge Perceptual Differences Between The Test Events  The impossible 1800 event lasted longer than the 1200 event, and there was more screen movement  Longer event may attract longer looking Rivera et al, 1999, Dev Psychol, 35, 427-435 It’s Preference For Greater Movement No Initial Habituation to Drawbridge Mean Looking time (sec) 30 * * Impossible 20 Possible 10 0 Block No block Rivera et al, 1999, Dev Psychol, 35, 427-435 But Baillargeon’s Control Condition Also Had The Same Movements?  The initial habituation event in Baillargeon’s study had the screen move 1800  Rivera et al. showed that infants have a spontaneous preference for the 1800 than the 1200 event  Habituation may have reduced interest in the 1800 test event, resulting in equal looking in Baillargeon’s control condition Rivera et al, 1999, Dev Psychol, 35, 427-435  Test Events Must Have The Same Appearance  The two test events had a different appearance  More change in the impossible event  More movement in the impossible event  Greater duration in the impossible event  Longer looking at the impossible event could be driven by these perceptual differences  Nothing to do with detecting the impossible event  How Could We Improve The Drawbridge Study?  Make the two test events look identical  Same amount of change  Same amount of movement  Same duration Possible 1200 Event Impossible 1800 Event A Modified Drawbridge Study  Hard and soft objects  Habituation event – soft object  Two test events  Possible – soft object  Impossible – hard object  Same amount of change  Same amount of movement  Same duration  Objects In The Events Sponge Wool Wooden Block Better Evidence: Hard & Soft Objects Habituation 1570 soft object A Same amount of change Impossible Same amount of hard object screen movement Possible soft object B Baillargeon, 1987, Cog Dev, 2, 179-200 Better Evidence: Hard & Soft Objects Same amount of change, same screen movement, but longer looking at the impossible event Baillargeon, 1987, Cog Dev, 2, 179-200  Initial Habituation Is A Problem  Habituation may have reduced infants’ interest in soft-object events  Infants looked longer at the hard-object test event because it involved properties  Result could be due to a novelty preference, not a violation of a physical property Initial Familiarisation Is Unnecessary  What influences infants’ looking during the test events?  What they already know?  What happened during familiarisation?  If infants have relevant physical knowledge, they should detect the difference between possible and impossible events without any prior familiarisation Haith, 1998, Inf Beh & Dev, 21, 167-179; Hood, 2001, Perception, 30, 1281-1284  Initial Familiarisation May Influence Looking At Test Events  Familiarisation trials may interfere with looking at the test events by inducing  Familiarity preference  Novelty preference  No preference  Baillargeon (1987) - habituation with the initial soft object may induce a preference for looking longer at the hard object (impossible event) Wang et al, 2004, Cognition, 93, 167-198 Houston-Price & Nakai, 2004, Inf Child Dev, 13, 341-348 Cohen, 2004, Inf Child Dev, 13, 349-352; Hood, 2001, Perception, 30, 1281-1284 Drawbridge Study – A Familiarity Preference?  Infants looked longer at the impossible 1800 test event which more closely resembled the habituation event  Could this longer looking also indicate a familiarity preference due to incomplete encoding? Schilling, 2000, Infancy, 1, 387-388  Habituation Trials at 4 Months  Condition A:  7 trials with 1800 event (Baillargeon method, familiarised to impossible event movement)  Condition B:  12 trials with 1800 event – (more trials than Baillargeon)  Condition C:  7 trials with 1120 event (familiarised to possible event movement)  Condition D:  7 trials with 1800 event, 6 month infants (faster processors) Schilling, 2000, Infancy, 1, 387-388  It’s a Familiarity Preference! Familiarity Preference 16 4 months 14 12 * 6 months Looking Time (sec) 10 180 impossible event ns 8 * 112 possible event 6 * 4 2 0 A B C D 7 Trials 12 Trials 7 Trials 7 Trials Schilling, 2000, Infancy, 1800 1800 1120 1800 1, 387-388  Initial Habituation Affects Results  Drawbridge study – Baillargeon included both habituators and non-habituators  Cashon & Cohen (2000)  Habituators looked longer at the possible event  Non-habituators looked longer at the impossible event – a familiarity preference?  Schilling (2000)  Clear evidence for both familiarity and novelty preferences, depending on amount of familiarisation Cohen, 2004, Inf Child Dev, 13, 349-352 Cashon & Cohen, 2000, Infancy, 1, 429-446 Schilling, 2000, Infancy, 1, 387-388  Is It A Familiarity Preference? Initial Display brief exposure, Test Displays incomplete processing 1+1=2 Look longer at familiar 2-1=1 Look longer Wynn, 1992, Nature, 358, 749-750 at familiar Cohen & Marks, 2002, Dev Sci, 5, 186-201 Initial Familiarisation Affects Preferences in Addition Task  5-month-old infants familiarised for 8 trials with  1 doll  2 dolls  Then shown an addition event (1+1)  Tested on  1 + 1 = 2 (possible)  1 + 1 = 1 (impossible) Clearfield & Westfahl, 2006, J Cog Dev, 7, 27-43  Infants Showed Novelty Preferences in Addition Task Familiarised to 1 doll Familiarised to 2 dolls 2 puppets 1 puppet 2 puppets 1 puppet Clearfield & Westfahl, 2006, J Cog Dev, 7, 27-43  Initial Familiarisation Affects Preferences in Addition Task  “A few exposures to a stationary doll or dolls can cause infants to look longer at any kind of display, regardless of its possibility or impossibility”  “Researchers must address issues relating to familiarization and other basic perceptual processes before assuming more complex understandings or knowledge in infants” Clearfield & Westfahl, 2006, J Cog Dev, 7, 27-43 Improving The Hard & Soft Object Study No habituation, only two test events Possible Impossible soft object hard object Predict longer looking at the impossible event, but this experiment has not yet been done  Improving VoE Procedure  No initial habituation phase  Eliminate familiarity/novelty preferences  Have a long delay between the initial event and the test events  Reduces the influence of initial event  Also a good test that infants have a representation of hidden object  Identical appearance of test events Reasoning About Hidden Objects Thin box event (possible) No initial habituation but some familiarisation 5 month infants tested after delay of 4 min Luo et al, 2003, Cognition, 88, B23-B32 Reasoning About Hidden Objects Thick box event (impossible) No initial habituation but some familiarisation 5 month infants tested after delay of 4 min Luo et al, 2003, Cognition, 88, B23-B32  Reasoning About Hidden Objects 5 month infants Looking Time (sec)  “Remembered the thick or thin box behind the screen 50 during the delay” 40  “Realised that the cylinder Thick block could only pass behind the 30 Thin thin box” 20 block  “Were surprised when this expectation was violated” 10 0 4 min delay Luo et al, 2003, Cognition, 88, B23-B32 VoE With Test Trials Only ? ? Wang et al, 2004, Cognition, 93, 167-198 VoE With Test Trials Only Wang et al, 2004, Cognition, 93, 167-198  VoE With Test Trials Only  4-month infants  Looked longer when a wide object was hidden behind a narrow occluder  Infants detected impossible events with no habituation or familiarisation trials Wang et al, 2004, Cognition, 93, 167-198 Interpreting VoE Results  What Do VoE Studies Tell Us About Infant Cognition?  Researchers such as Baillargeon and Spelke have made sweeping claims about infants’ abilities based on the results of VoE experiments  These include:  Knowledge of physical events (e.g. properties of solid objects)  Concepts such as solidity, continuity, causality  Ability to form detailed mental representations  Abilities at reasoning about events  Critics point to the limited information which VoE methods can provide  And they caution against the tendency to over-interpret the findings  Haith’s List of Outright Psychological Felonies  Use of a minimalist perception paradigm that taps only a limited aspect of knowledge  Competencies, not processes Haith, 1998, Inf Beh & Dev, 21, 167-179  Minimalist Perception Paradigm  Habituation and the visual comparison paradigm can answer only Yes/No questions:  Can infants discriminate shapes?  Do infants recognise a stimulus?  Habituation paradigm was developed to discover competencies, not the processes underlying those competencies  VoE paradigm encourages dichotomous answers to complex questions:  Can infants do arithmetic? – Yes/No Haith, 1998, Inf Beh & Dev, 21, 167-179; Hood, 2001, Perception, 30, 1281-1284 Recognition Of Familiar Stimulus  Habituation paradigm shows Habituation that infants recognise a stimulus  But what do they recognise? Is it the entire face, or just a feature such as hair colour?  A difference in looking time indicates memory for the Novelty Preference familiar face  But provides no information about what has been encoded Hood, 2001, Perception, 30, 1281-1284  Pyramid of Methods Paulus 2022, Inf Child Dev, 31, e2306, https://doi.org/10.1002/icd.2306 Both Adults And Children Can Detect A Difference Without Knowing What Causes That Difference Something Different Here, But What Is It? You Can Spot A Difference Without Knowing What It Is Thompson, 1980, Perception, 9, 483-484  Longer Looking At An Impossible Event Without Knowing Why  There is a difference between detecting a violation and knowing the reason for the violation  Both adults and children readily detect incorrect movements of an object when watching an event  Both make errors when asked to predict its movement Hood, 2004, Dev Sci, 7, 415-417 How Does The Load Fall? A B C D E Chosen by Correct many adults The Curved Tube Problem Kaiser et al, 1985, J Exp Psychol Mem & Cog, 11, 795-803  Expectations and Knowledge May Differ  Adults used their knowledge to choose which outcome they thought would occur  They frequently chose the wrong outcome  Then they judged which simulations looked correct  They were much better at detecting the correct simulated outcome  Looking data do not provide complete information about knowledge Hood, 2004, Dev Sci, 7, 415-417 Smart Infants And Dumb Toddlers  Haith’s List of Outright Psychological Felonies  Indications for earliest fragments of a concept used as evidence for virtual mastery of the concept  Assertions that young infants know things far earlier than seems reasonable Haith, 1998, Inf Beh & Dev, 21, 167-179  Smart Infants And Dumb Toddlers  Studies showing apparent lack of knowledge of solidity in 2-year-olds  Hood (1995)  Berthier et al (2000)  Hood et al (2000)  Hood et al (2003)  Are Solidity & Continuity Understood at 3 Months? Where the ball comes to rest in each condition Habituation Consistent Inconsistent Hand holds ball above screen and then drops it Spelke et al, 1992, Psych Rev, 99, 605-632 Where Is The Frog? Hood et al, 2000, Child Dev, 71, 1540-1554 Where Is The Frog? Hood et al, 2000, Child Dev, 71, 1540-1554 Knowledge Of Solidity Did Not Guide Search At 2 Years Correct Correct Search Location Shelf: trial 1 45% Upper Shelf: trial 2 55% Upper No shelf: trial 1 50% Lower No shelf: trial 2 30% Lower Hood et al, 2000, Child Dev, 71, 1540-1554  2-month-olds Look Longer At The Impossible Event  Infants understand  Solid objects cannot pass through each other  Objects move on a continuous path Spelke et al, 1992, Psych Rev, 99, 605-632 Can 2-Year-Olds Find The Ball? Barrier Screen Doors Ramp Berthier et al, 2000, Dev Psychol, 36, 394-401 Success! Keen, 2003, Current Dir Psychol Sci, 12, 79-83 2-Year-Olds Can’t Find The Ball! Barrier Screen Doors Ramp % Correct First Searches 2 yr 2.5 yr 3 yr 22% 34% 74% Berthier et al., 2000, Dev Psychol, 36, 394-401 Looking Versus Searching Toddlers look correctly, but search incorrectly Hood et al, 2003, Dev Psychol, 39, 61-7  Looking Versus Searching  2.5 yr-old children made many errors on the search task  But the same children looked longer at an impossible outcome with the same apparatus  They detected an incorrect (impossible) outcome  But were unable to search correctly themselves  Performance on the two tasks was not correlated Hood et al, 2003, Dev Psychol, 39, 61-70  A Caution As Haith pointed out, we must be careful not to infer almost complete mastery of a concept from the earliest signs of its emergence  Why The Discrepancies Between Infants And Toddlers?  VoE requires only looking  Infants never have to make something happen  VoE involves evaluation of an outcome  No need for prediction  Infants can work backwards from the endpoint, and all the relevant information is given  Search involves prediction of an outcome  No endpoint from which to work backward, so relevant information has to be selected Willatts, 1997, Couch Potato Chapter Keen, 2003, Current Dir Psychol Sci, 12, 79-83  Violation of Expectation or Experience?  VoE may not involve prediction and expectation  May just be detection of a novel or odd event (e.g. an object floating in mid-air)  Violation of Experience rather than Expectation would be more appropriate  Detection of impossible events – could provide a building block for constructing further knowledge Willatts, 1997, Couch Potato Chapter Keen, 2003, Current Dir Psychol Sci, 12, 79-83 Rich Interpretation  Haith’s List of Outright Psychological Felonies  Over-interpretation of findings as evidence for high-level cognitive operations  ‘Belief’, ‘representation’, ‘surprise’, ‘reasoning’  This is “rich interpretation” Haith, 1998, Inf Beh & Dev, 21, 167-179  Rich Interpretation “How much of cognition is in the head of the infant, and how much in the mind of the theoretician?” Haith & Benson, 1998, Chapter The Carrot Study Baillargeon & Graber, 1987, Cog Dev, 2, 375-392  Rich Interpretation These were Baillargeon & Graber’s conclusions: “Infants (a) realised that each carrot continued to exist, (b) assumed that each carrot retained its height, (c) believed that each carrot pursued its trajectory behind the screen, and therefore (d) expected the tall carrot to be visible in the window” Baillargeon & DeVos, 1991, Child Dev, 62, 1227-1246 No Anchoring Of These Higher Level Constructs  No behavioural indicators of realisation, assumption, belief, inference, reasoning, expectation, categorisation, surprise, symbolic representation  These constructs can be measured in older children  Using the same terms in infancy implies a connection that we cannot prove Haith, 1998, Inf Beh & Dev, 21, 167-179  Baillargeon’s Response  “Evidence that infants look reliably longer at the unexpected than the expected event is taken to indicate that they:  Possess the expectation  Detect the violation  Are surprised by the violation  ‘Surprised’ is used simply as a short-hand descriptor to denote a state of heightened interest or attention”  So why not refer to ‘interest’ or ‘attention’, instead of ‘surprise’? Baillargeon et al, 2006  The Limits Of Looking Time “The apparent dissociation of duration measures from search performance gravely undermines the position of proponents of a rich interpretation of the looking time paradigm” Looking time differences tell you that children noticed a difference, but do not tell you anything about how they did it Hood et al, 2003, Dev Psychol, 39, 61-70 Can We Improve VoE?  Is It Time To Leave The VoE Method Behind?  This method relies on too many questionable assumptions  Its conceptual foundations are thin  Its findings are empirically not conclusive  “Theoretical inferences based on VoE findings are thus too speculative to allow for solid progress in developmental psychology” Paulus 2022, Inf Child Dev, 31, e2306, https://doi.org/10.1002/icd.2306  Include Other Methods To Provide Converging Evidence  Wang & Kohne (2007) used a manual search task to show that priming allowed infants to identify height as a critical variable  Their results supported a similar VoE study (Wang & Baillargeon, 2005) Wang & Kohne, 2007, Dev Psychol, 43, 1513-1522 Observing The Unexpected Enhances Infants’ Learning And Exploration Stahl & Feigenson, 2015, Science, 348, 91-94 Observing The Unexpected Enhances Infants’ Learning And Exploration  “Infants learned more effectively about objects that committed violations, explored those objects more, and engaged in hypothesis- testing behaviors that reflected the particular kind of violation seen”  Let infants explore the objects involved in impossible events to confirm a violation of expectation  Interesting to do this with toddlers (Studies by Berthier et al. 2000; Hood et al. 2003) Stahl & Feigenson, 2015, Science, 348, 91-94  Pupil Dilation – A Possible Measure of Expectation & Surprise  Pupil dilation has been successfully used to detect surprise in adults  Kruger et al. (2020) showed pictures of animals combined with matching or mismatching animal sounds  Participants were aged 1, 2, 3, 4, 5, 6 yrs & adults  All ages showed greater pupil dilation for mismatches than matches Kruger et al., 2020, Child Dev, 91, 2221-2236 Pupil Dilation May Not Be Useful With Younger Infants Toy moves into window and either remains or leaves. Door closes and then opens to reveal either the toy or no toy. Patzold & Liszkowski, 2020, PlosONE, 15(4): e0230913  No Effect With 10-Month-Olds  18-month-olds showed greater pupil dilation when expectations were violated  10-month-olds showed no difference in pupil dilation – this may not be a sensitive measure with younger infants  Patzold & Liszkowski offered an odd explanation: “a possible interpretation is then that young infants do not automatically represent occluded objects”  But 8-month-olds do search intentionally for hidden objects, and do not search when nothing is under a cover Patzold & Liszkowski, 2020, PlosONE, 15(4): e0230913  One Final Study Social Looking – A New Measure  Social Looking – Another Possible Measure  6-month-olds were habituated to an event in which a hand placed a toy behind a screen, and the screen was then lowered to reveal the toy  On test trials, a novel toy was placed behind the screen, and the screen was lowered to reveal  The same novel toy (novel trials)  The original familiar toy (VoE)  Would infants detect the violation? Dunn & Bremner, 2017, Dev Sci, 20: e12452 Habituation Tests  Social Looking – Another Possible Measure  There was no difference in total looking time on the novel and VoE trials  But infants looked more often at their parent after the VoE trials – increased social looking  Increased social looking occurred even though infants saw the familiar toy  Social looking is a response to VoE over and above any response to novelty Dunn & Bremner, 2017, Dev Sci, 20: e12452  Summary - Problems with VoE  Too simple a method to reveal much about infant knowledge  Is it violation of an expectation or simply detection of an odd or novel event?  Longer looking at the impossible event indicates that infants see a difference  VoE paradigm does not provide information about why infants see the events as different  Encourages rich interpretation – excessive interpretation of findings  Haith’s initial concerns have yet to be met (Tafeshi et al., 2014, Hum Dev, 57, 222-240) Haith, 1998, Inf Beh & Dev, 21, 167-179 Summary – Interpreting the Results  Noticing a difference is important because infants may have some knowledge of the physical principle  But the earliest fragment of a concept is not evidence for mastery of the concept  Performance of ‘dumb’ toddlers is striking evidence of this fact  Improving the VoE Paradigm  Remove the habituation/familiarisation phase  Test events must look exactly the same  Maybe add a delay between occlusion and test events to ensure a representation is stored  Be cautious with the interpretation  Try and obtain converging evidence using other methods (e.g. reaching)  Include different measures, such as social looking