Reading Flashcards
what is reading
the active construction of meaning: the mental representation of the situation described by the text
requires one to identify the words, activate contextually appropriate meanings, form causal connections with the text and prior knowledge to make inferences about a situation.
what are the challenges to reading
need to link visual representations of letters in a word to their sound and meanings
Complex, dynamic, cognitively demanding yet feels effortlessly easy for skilled readers
how is the reading system flexible and generalisable
we are able to read non-words that follow rules
can infer meaning of unfamiliar words based on their roots
we can distinguish the meaning of words with the same orthography but different phonology
we can distinguish the meaning of words with different orthographies but the same phonology
how many words on average can skilled readers read (Brysbaert, 2019)
238 words per minute
how can we explain the normal distribution curve of reading variability
Biologically secondary skill that’s parasitic on more primary systems
high heritability of learning skill with clear neural correlate despite reading systems being artificial inventions that are culture dependent and relatively modern. unnatural act that needs to be taught.
this suggests there has not been selection for reading genes, rather that other genetic factors affect or ability to read: language, perception, memory and learning skills.
what is Seidenberg’s triangle model of reading
theoretical framework to understand learning
Reading is modelled by representations across 3 different levels: orthography, phonology & semantics with weighted connections (representations)
P-S mappings Arbitrary, meaning not predictable from sound (for morphologically simple words)
O-P mappings Systematic and quasi-regular, spelling predictable from sound and vice versa (quidditch)
O-S mappings Direct connections from spelling to meaning (sail-sale)
what mapping is learnt first in alphabetic languages
Orthography to phonology
i.e. children need to learn the alphabetic principle
what is the alphabetic principle
the fundamental insight that the 26 letters of the alphabet work together in systematic ways to represent the sounds of the language
why can’t we use paired associate learning
doesn’t give us the ability to generalise
need a rule based system
how do inconsistencies in mappings affect learning
english has a high number of inconsistencies such that after 1 year of teaching children can read 41% of words while Italian children can read 92% as their language is highly regular
(Willingham et al., 2017)
confounds age in first year of school and quality of teaching
do children have explicit knowledge of sound structure
Awareness tasks require children to make explicit judgements about the sounds of words (or nonwords) or to perform manipulations of the sound structure of SPOKEN words – Counting, deletion, blending tasks (Melby-Lervag et al., 2012)
phonological awareness predicts later reading ability
what evidence is there to support the link between phonological awareness and reading ability
cross-sectional data
(liberman & Shankweller, 1974) asked children to clap out number of syllables or phonemes in a word (phoneme: 0% age 4; 15% age 5; 70% age 6)
meta-analysis (Melby-lervag et al., 2012) children who find reading difficult (e.g with dyslexia) also struggle with phonological awareness r=.57
longitudinal data (muter et al., 2004) early skills in phonological awareness age 4 predicts reading ability 1-2 years later
training data (Hatcher et al., 1994) training in phonological awareness and reading combined the only condition to improve outcomes
conc: Something about PA is related to the early process of learning to read
Once children have alphabetical principle children have the fundamental insight to how letters work together to represent language via sound
why should we question the sharp divide between the logographic and alphabetic stages of reading
Reading non-words relies on generalising from knowledge of how reading system work (which we can do)
phonetic cue reading (Rack et al., 1994) logographic children taught to associate abbreviations with pronunciations showed recognition was better for words with phonetic cues compared to controls
what is Frith’s (1985) Stage Theory of reading
logographic –> alphabetic –> orthographic
logographic: builds slight vocab on basis of visual appearance, no attention to individual letters, can only read known words e.g Masonheimer et al., 1984 removing logos stops children being able to ‘read’ words / partial visual cues: its the long one
alphabetic: pays attention to individual letters, learns to decode, can attempt to pronounce nonwords, regularisation errors
orthographic: develops orthographic strategy, sensitivity to larger units that GPCs, notion of automatic word recognition
what is orthographic knowledge
the conventions, regularities and quasi regularities that characterise each writing system
What is Share’s 1995 Self-Teaching Hypothesis of Orthographic Learning
an alternative to stage models
Phonological skills are essential and fundamental
Allow children to decode or recode the printed form of a word
To translate the printed form of a word into its spoken form
Needs to be successful because it provides two key opportunities
To acquire word-specific orthographic information
Accumulate statistics about how the language words
what support is there for the self-teaching hypothesis
Orthographic learning via self-teaching observed in 2nd grade children learning to read Hebrew (Share, 1999, 2004)
Children learn something specific about the form ‘yait’ after few exposures: and general too
if the phonological aspecits of decoding are prevented from happening kids don’t show orthographic learning
Reliable and equivalent learning following 1, 2 or 4 exposures in Hebrew (Share, 2004)
Reliable and equivalent learning at 3-day, 7-days and 30-day post-tests in Hebrew (Share, 2004) - durable
Evidence for rapid and durable single trial learning in 3rd grade children reading Hebrew
(retesting in english speaking found some consistencies across languages, but some differences
Exposure effect although this is moderated by a frequency effect
Durability effect although there is some decline at post-test)
How can we learn to read through statistical learning
Product of acquiring orthographic knowledge akinto a statistical learning approach
Rather than a series of qualitatively distinct stages we acquire orthographic knowledge over time
Phonological skills essential and fundamental
Decoding provides an opportunity to translate a printed word into its spoken form
This successful (but potentially laborious) decoding attempt provides an opportunity to acquire word-specific orthographic information
And accumulate statistics about how the language works (statistical learning)
what does staecy et al., 2019 propose as a statistical model for reading
initial reading is orthographic –> phonological –> semantic (triangle) which is dependent on the alphabetic principle (becomes o-> s with expertise)
this plus reading experience allows children to internalise probabilities, regularities and quasi-regularities and generalise this knowledge to new words
what are connectionist models of learning to read
No rules and no stages
Simple associative processes learn to map distributed inputs na outputs the triangle model learns well and captures behavioural data well
top-down: context -> meaning / meaing-> orthography and phonology
bottom-up: orthography and or phonolotgy> meaning
conclusions of reading
Reading is a learned skill but one that depends on language, perception, memory, learning…
the alphabetic principle at the foundation (in alphabetic systems)
phonological skills are critical
children move from something that looks logographic to skilled, but is this best characterised by stages or, are the same mechanisms involved with experience providing input that drives learning?
