Science of Reading Words and Dyslexia Part 1-B Charles Arthur Both Sides of the Theory and Its Expanded Application. 5/8/18 1 Part 1-B, Day Two: Both Sides of the Theory: Speech and Print From the Side of Speech (Pg. 2-7) We read with our eyes, but the starting point for reading is speech. Speech: made possible by a large volume of words, orally learned, and held in memory. (evolution of brain size) Pre-spoken words, as bundles of alphabetically arranged phonemes, composed from 43 phonemes in English, form thousands of words. Bundles, to be cognitively tapped into, at the phonemic level by printed words for reading. Good readers must posses a large storage of vocabulary that are accessed, tapped into, at the most basic level. 5/8/18 2 1
Accessing Spoken Vocabulary from Print can be Difficult Letters become matched and stuck to phonemes, enabling the correct words to be found. ( piggy backing reading onto speech) For speech, the brain automatically manages and forms the phonemes, received or produced, into words. nothing done by the speaker or listener. (see Pt II) For reading, the phonemes need a stronger identification than needed for speech, In order to form the match with designated letters, To create a written language that is easier to read than logo-graphic symbols. Gives the phonology of reading prominence over print. We can read, an alphabetic language, because we can speak. (unknown by the original invertors.) 5/8/18 3 Here s the rub in making the link (p. 5-7) Difficulty Finding the Phonemes in Speech Finding and identifying the individual phonemes for L/S matching is not a natural act. It must be learned. It s not needed as much for most speech. Most people have some difficulty. Approx. 30% more so. The skills with phonemes seems to be distributed along the normal curve. - low, average, high sensitivity low (dyslexia); average (elusive); high (easy) Printed words are structured slightly different than speech In neat rows, lines, not in bundles. (see page. 6) An alphabetic print can only approximate speech The ear needs tuning-up to hear and identify phonemes in the pronunciations. Readers need to have knowledge of the patterns and structure in the oral form. explicitly vocalizing all pieces in words.- learned from careful listening, before making the matched. 5/8/18 4 2
Two ways an alphabet can represent phonemes in spoken words. Approximate linier arrangement, but readable Four distinct sounds or phonemes Stop Actual bundles, but unreadable. Overlapping sounds or phonemes. 5/8/18 5 What is required from the speech side for reading? Knowledge of the hidden alphabetic structure of spoken words, - a consensus on causal relationship to learning to read. This means hearing and articulating individual phonemes in their arranged order in words, not a natural by-product of speech. the ear needs additional tuning and training. An increased demand for this knowledge comes from reading an alphabetic writing system, beyond what is needed for speech. This is the root source and cause of Dyslexia. 5/8/18 6 3
How critical is this knowledge? The discovery and documentation of phonemic awareness is considered the single most powerful advance in the science of pedagogy of reading this century. -Marilyn Adams This discovery solved the instructional impasse found in teaching alphabetic knowledge, phonics, That was difficult for so many children, especially those with a learning disabilities like dyslexia. - the source of the disability This critical element of reading has been evaded by many programs, phonics or not. With ill affects on at least 25% 5/8/18 7 The Learning Disability of Dyslexia Because of what is now known about reading, we know that any weaknesses with the phonology of language will have serious effects on learning to read words. The lower end the normal distribution of this natural capability is the precise cause of dyslexia. This indicates a weakness with the phonemes within spoken words --- finding (hearing) and articulating (vocalizing). This results in difficulties with making and learning letter/sound matches, required for cipher sight-word reading.. This weakness also has ill effects on short term memory, rapid oral naming response, as well as other skills needed for decoding words and listening comprehension. (pg. 7,8) A phonological weakness can be corrected. 5/8/18 8 4
CONSEQUENCES The weakness with phonology of language is more apparent in reading, writing and spelling, than in speech, due to the added demand on the phonology of speech. Although dyslexia can often be detected in speech. This weakness can be corrected if caught early, And, if PA exercises and letter/sound connections are taught carefully and directly by teaching methods designed to accomplish this. (direct instruction implications, see Part III) The term dyslexia might fade into obsolescence like smallpox, polio or pertusis. See Part Ib, pg. 8 for graphics, definition, behavioral affects and possible prognosis. 5/8/18 9 From the Side of Print (p. 9-11) An invention that had HUGE consequences, that took over a 1000 years to form documents Printed words have a distinct advantage over oral. they can be put on pause, saved and repeated. An alphabetic writing system is a lot easier to learn and can help in making and strengthening phonemic sensitivities and skills. speech and reading are a two-way street In spite of its prominence, print is secondary... the alphabetic structure of (spoken) words is not to be found at the surface of the acoustic signal but only at a deeper, less accessible level. We read with our eye, but the starting point for reading is speech. We can read an alphabetic language because we can speak. 5/8/18 10 5
A two-way street - some direct instructional implications Learning to read an alphabetic written language, itself, forces a certain amount of phonemic awareness and sensitivity Both sides can be learned in the same beginning lessons. But separated at first, with and without letters. They tend to reinforce each other in memory And, in the act of reading words, the ears and voice The ear directs eye movement, attention and coordination. Instruction in phonemic skills and sensitivity that leads into the L/S sounding-out of letters in words, makes learning the connections in decoding easier. (facilitating and increasing the chances for success). How this lead-up is taught is crucial. Segmenting vs continuous sounds of phonemes A critical difference of view among researchers. 5/8/18 More on this in Part III, Instructional Implications 11 Writing is not language, but merely a way of recording language by visible marks. The Lexical Quality Hypothesis (LQH). Calls for learning a high quality of print for the two-way process to work. how thoroughly do readers need to learn the details of print in order to make firm L/S bonding? How trained should the eyes of readers be to notice the details of print? How does this affect bonding? How the letters are used is the central theoretical question of word identification. While all letters must be perceived, not all need to be seen equally. Some get a relatively high level of activation (attention) and others a relatively low level from the word. (used just visually or phonetically connected?) When people read text, the print fills their minds with ideas. The route to these ideas begins with individual printed words. Eye movement studies show that when readers read a text, their eyes land on practically every word See quotes by Perfetti and Ehri on p. 11 5/8/18 12 6
These characteristics allow word reading and meaning retrieval to be rapid and relatively automatic given a familiar printed word. High Quality Visual Representations from. Increased repetitions (from?) Increased quality of detail (how? Detailed knowledge of letters strengthens phonemic sensitivity. (why?) word-form knowledge is critical in skilled reading. Phonological training increased sensitivity of the visual word form to the phonological structure of words. p. 27 (a two-way path) The need for phonemic sensitivity extends beyond what is needed for speech. 5/8/18 13 The Theory s Expanded Application How alphabetic knowledge is applied to reading a large volume of variable words, in texts. Especially with opaque English spellings. p. 12 Letters representing more than one phoneme, C G Y Phonemes that have alternative spellings. vowels, 48% regular vs. 94% in French Contrasted spellings, silent (non-phonetic) letters 250+ common, rule-based L/S correspondences Numerous and varied but still regular, and quasi-regular, only partially regular With some grammatical reasons. p. 28-29 Good readers are able to read all kinds of words equally well. 5/8/18 14 7
An Analysis: How varied are they really? Gough: rules apply for the majority of English words. Out of all the hundreds possible, from 200-500, 55 common L/S rules can be applied to about 80% of all words. Regular spellings: both simple and advanced Mostly consistent, some advanced. The Other 20% Rules? Unusual variations are a matter of degree and can be arranged in teaching along points on a continuum. Basic alphabetic decoding skills, as a foundation goes a long way. Limitations to Ehri s theory? From learning common letter patterns or word families. Two ways of learning to read words: Memory, for irregular, and Decoding, for regular? 5/8/18 15 Quasi-regular Words From a Computerized Model p. 14-15 Words with variations are learned with the use of alphabetic knowledge, which extends the use of Ehri s theory. Thus illustrating that two methods, rote memory and decoding, are not necessary, or commonly used. New words learned on basis of previously learned patterns Another place where theory directly implies instruction. In describing how the alphabetic principle is extended beyond the basic 55 of 80% on a carefully sequenced continuum 1. New and unusual phonetically based words 2. Use Alphabetic principle as a base for quasi-regulars 3. Use patterns that have already been used for new variations. Can apply to thousands of words, Yet daunting.. 5/8/18 16 8
An analysis of word frequency can help From a 5 million word sample: Most Frequent: more functional, less meaning bearing. 50% came from a pool of 100 words, (78 regular). 25% came from a pool of 900 words, (80% regular?) (75% came from a pool of 1000 words.) Least Frequent: (some of the 80% GPC may be here.) Mostly longer, more complicated words, that carry more meaning. 15% came from a pool of 4000 words 10% came from a pool of 80,000 words -Fluency and accuracy are more challenging. All requiring fluent reading. (still daunting) 5/8/18 17 Is the GPC theory adequate? Given the variations, the subtle nuances in speech, the approximations of the alphabet, the total volume, variations and word frequencies. Gough sees a need for additional knowledge of unique spellings to be used as needed built on top of a GPC theoretical foundation. Would this slow the reading down some? It typically doesn t Is there something more at play to make the GPC theory work, at the automatic level, on a grand scale with all these variables? Ehri is uncertain. P. 16 5/8/18 18 9
Is there something more that makes it all work? Gough implies that there is something more by referring to cipher reading as a metaphor. P.17 But doesn t know what it is. How this daunting task is done still contains mystery. Ehri is uncertain about the extent of her theory within the complex English system. Research is formidable in support of the theory, but there are. Some lingering questions (see Extension A) 5/8/18 19 From Experimental Behavioral Research HOW DO WE KNOW? Understanding reading words as phonologically based is almost universal, even before brain imaging. Ehri reviews many, Perfetti summarizes. P.18 Seidenberg lists a number of reasons. The extent of the GPC theory, in explaining reading words in all circumstances and on a large scale, based on behavior al research is still questioned. What is missing? See What remains to be explained? P.20 Brain imaging confirms, clarifies and illuminates. 5/8/18 20 10