Debunking the Types of Dyslexia: One Cause, Many Myths
This article by Dr Piotr Wozniak is part of SuperMemo Guru series on memory, learning, creativity, and problem solving.
Educational dyslexia
Dyslexia is a great illustration of how "science" of education can turn adult anxieties into a hell for children. Parental obsessions about the performance of their kids fuels an industry of semi-competent helpers, vested interest and an army of lemmings and parrots in peer-review science motivated by publish or perish.
In healthy children, there is basically one type of dyslexia that you need to know about: educational dyslexia. Educational dyslexia is born at school under coercion, and its mechanism is very easy to understand and prove (see experiment below). However, what is simple and obvious may still struggle against the mythology generated by the dyslexia industry (see: Dyslexia Conspiracy).
Strengths vs. weaknesses
All kids differ. Their brains differ wildly because the main job of the brain is to adapt to the world. In identical environments, a single toy or a book on a shelf will change a child into an entirely different person. Poor understanding of that variability, and the wish to mass produce little intellectuals result in developing "norms" that are the root of all evil.
The two main axes of variability that result in the emergence of dyslexia are: (1) different preferences stemming from different adaptation trajectories, and (2) different developmental trajectories. The difference in preferences may be as simple as the love of comic books as opposed to the joy of watching YouTube. Difference in trajectories can be seen in developmental milestones. A genius brain is often slow to develop, while the popular perception is the opposite: kids who are first to walk, speak or remember are considered smartest (see: Precocity paradox).
A popular explanation of educational dyslexia is that kids who are late with a skill are targets of intense "instruction". They are scared into feeling "worse". "Teaching" identifies areas of "weakness" and those areas are subject to most rigorous training. If the kid's natural development does not let him break out from this feedback loop, he may me marked for life as a dyslexic.
Dyslexia is just one of many "conditions" that results from educational norm setting. Dyscalculia, dysgraphia, ADHD, or speech problems may also emerge in similar settings.
The mechanism is always the same: (1) spot the weakness, (2) focus on the weakness, and (3) amplify the weakness through rigorous exercise under stressful coercions.
Education should always amplify strengths! I might say: (1) spot the strength and (2) amplify it. However, this process occurs naturally. Kids find their own strengths by pursuing their passions.
In that light, education should rather focus on not disrupting the natural process. Its main role should be protection, educational enrichment, and inspiration.
Learning to read
The right way to read is to achieve one's goals by overcoming obstacles posed by print. For example, a child may enjoy computer games. The first word he might learn to read is "Play". The first letter might then be "P". There may be months before he learns to read "Stop" or "Level" and then letters "S" and "L" or perhaps the first letter of one's own name. In team games, conversation and print obstacles interleave, while peers focus on one thing only: playing. Unlike teachers, they do not correct you. The play itself provides penalties if you misread a message or mis-pronounce a command. Friends will correct you only when you err gravely or if they get annoyed with the same mistake repeated a hundred times. They will not drill you nor stress you. Those who are not nice, unlike bad teachers, will instantly be substituted! The kid can always play with someone else. Overtime, with no stress, no pressure and with the pleasure of his own achievement, the kid starts to read. His reading will be of highest quality: full focus on comprehension with spelling or pronunciation playing the second fiddle. See: Don't teach your child to read
Inducibility of dyslexia
It can be demonstrated by a simple experiment that dyslexia is inducible in anyone. You do not need a faulty gene nor neurodevelopmental deficit. All you need is to (1) start early enough and (2) exert sufficient coercive pressure!
The experiment below occurred serendipitously. It should never be repeated. In retrospect, we can say that an attempt at replication would be unethical. Parents may use coercion with the best intent in their heart. I never discouraged parents from using SuperMemo for learning with little kids. I changed my mind when first data sets arrived. I could see that SuperMemo does not work for kids, and should never be used, unless voluntarily. The need to use spaced repetition may arrive only in 12-14 year olds. Even then, I suspect, school might be the main drive behind those early motivations.
The experiment below spanned over a decade and provides rock-solid proof for the existence of educational dyslexia.
Experiment
The following experiment proves a phonological toxic memory engram in a healthy brain in the process of inducing educational dyslexia.
A child aged 2 was asked to memorize the letters of the alphabet. In order to perpetuate the memory, SuperMemo was used with the intent to keep the entire alphabet in memory. In the process of spaced repetition, memory patterns are presented at time intervals that ensure a set level of memory retention. In the ideal case, for a well-formulated memory engram, an adult would need just a few repetitions in a lifetime. The same might be the case in children, however, their mnemonic capacity is greatly reduced, which makes short-term storage harder. Most of all, childhood amnesia makes it hard to form long-term memories. For this reason, spaced repetition should never be used with little children.
Proving a healthy circuit
In the experiment, the child showed perfect recall of letters M, A, S, and H from the age 2 till adolescence. Those letters were memorized at the age of 2 and never forgotten (scoring perfectly in between 35 and 62 repetitions in the course of a decade). This large number of repetitions indicates manual intervention with the duration of review intervals (i.e. not typical for spaced repetition), but it is also a strong indicator that a child received a solid reinforcement of the correct associations in the course of life.
Good recall indicates correct functioning of that part of the reading circuit that is responsible for the necessary pattern recognition and phonological processing. The circuits needed to associate the visual patters of letters on input with correct pronunciation (and motor system needed in articulation) were in place and operational. We might hypothesize that spaced repetition with no memory lapses was relatively harmless here.
Inducing toxic memory
In the same experiment, the same child with a healthy reading circuit, struggled to form a correct association for letters Y and U. From the very beginning, at the age of 2, the letters showed strong interference between the associations formed by the two visual patterns. In a total of 173+142=315 repetitions, 32+34=66 memory lapses were recorded. This means that each time the child was able to form a correct association, it was quickly erased via memory interference. If interference is specific, i.e. affect a specific pair of letters, we know that memory traces persist. They are simply not the traces the educator expects. It is not that Y is confused with A or S. It is always confused with U. That's a classic case of toxic memory that adults encounter frequently in the course of spaced repetition, esp. before developing healthy knowledge formulation habits (see: 20 rules of knowledge formulation).
The correct resolution to the problem of interference is forgetting. Instead of reinforcing wrong memory with drills, we should simply await a time when a new association forms and reinforce the new correct association. The situation is made worse with spaced repetition, which is great at long-term memory consolidation with little concern if the reinforced memory is made of desired associations. In the presented experiment, the process resolved at the age of 8. A school-age child is coerced into a multitude of contexts in which she has to interact with print. A good side effect of that interaction is an increased chance of forming correct associations that may overcome the interference generated in SuperMemo. Still, the correct course of action would be to delete wrong associations, and possibly return to completing the alphabet set at later time (if ever needed).
Toxic memory diagram
Figure: Toxic memory forms when learning penalties interfere with learning rewards and make it impossible to resolve conflicts in an interference zone of a pattern recognition network. The diagram shows a simple concept network illustrating the process of forming toxic memories. In the diagram, the brain attempts to associate a visual pattern of a letter "Y" with its correct pronunciation [waɪ]. In the course of learning, natural interference is a norm, and is quickly resolved by rewards of correct association signal on the output, and penalties of incorrect associations. If interference makes it impossible to establish a pattern in memory, the association is rejected and awaits subsequent opportunities to be re-established in a new configuration. However, in conditions of stress, the interference from penalties on output may prevent the resolution of conflicts in the interference zone. If the situation persists, e.g. as a result of frequent drills at school, forgetting will not be able to clean up the unwanted memory. The interfering pattern may persist. Instead of enhancing synaptic connections in the pattern recognition network, it may enhance negative associations leading to intrinsic penalties (stemming from dissatisfaction with learning), and extrinsic penalties (e.g. dissatisfaction of a teacher, peer ridicule, etc.). The loop of extrinsic penalties reinforcing the interfering pattern may be particularly harmful. This is how coercive instruction does harm to mental health of children. At later stages of this process, the pattern on input, instead of being resolved with the right association, generates an instant penalty signal. This occurs in educational dyslexia where all things associated with print and reading may bring up the state of anxiety or fear
Self-directed learning
Presented experiment shows inducibility of toxic memory and dyslexia in any healthy child. However, to complete the picture, we need to ponder how optimum learning should look like.
The same child voluntarily decided to learn Chinese at the age of 8. Again SuperMemo turned out to be useful in collecting the evidence. This time, it was not coercive learning. The experiment was short-lasting, but voluntary.
The child showed perfect recall of Chinese ideograms at the age of 8. In 149 repetitions of 28 symbols presented once, not a single lapse was recorded! Most of that learning took place in a 6-8 months period. The learning was based on the use of mnemonics which facilitate appropriate memory formation.
That stellar and surprising performance in Chinese comes from:
- superiority of self-directed learning over coercive learning
- the importance of "network readiness", which is a combination of biological maturity and the presence of the underlying semantic framework
- absence of a critical period in letter/ideogram visual pattern recognition
- the power of mnemonics that facilitates recall and consolidation
- the power of meaning (ideograms may provide reinforcement through meaning, e.g. "door", while letters of the alphabet such as "R" do not)
As there are no critical periods for recognizing letters, early alphabetic specialization is not necessary. Adults can learn to read in any language. Early academic instruction in reading is likely to cause more harm than good.
Educational dyslexia
Educational dyslexia is a condition in which a large number of atomic toxic memories accumulates in the area of reading skills. Those memories may be visual, auditory, phonological, orthographic, lexical, syntactic, articulatory, graphomotor, cross-modal, etc. Depending on the preponderance of individual memory sets, types of dyslexia may be delineated (visual, auditory, etc.).
The prime cause of educational dyslexia is coercion that perpetuates toxic memory engrams through repetition. In the course of instruction or therapy, a child is unable to rest through a sufficiently long refractory period. There isn't enough time for forgetting to do the necessary clean up job. The problem is the inability to ignore knowledge that does not stick to memory in a given context. Repeated exposure in the same context, due to coercion will result in perpetuating a memory (advancing through the stages of futile, persistent, and parasitic memories as defined in Toxic memory).
Types of dyslexia
Types of dyslexia stem from child's interests and preferences. If a kid is visual, he may be negligent about speech. He will be at an increased risk of phonological/auditory dyslexia. If a kid is a talker, she may love sounds, phonemes, and letters but might show weaknesses at visual processing. All those details will be meticulously noted by hawk-eyed "educators". Once weakness is spotted, it will become an excuse for further instructional torment. That's a straight path to educational dyslexia of one type or another (see: Don't teach your child to read).
Only connectionist models explain all forms of dyslexia, and all those models show dyslexia to be inducible in conditions of coercion. In connectionist models, all association routes involved in learning are built incrementally via interconnected feedback loops. For the learning process to proceed smoothly, self-directed control is essential.
Phonological dyslexia
If 80% of children with dyslexia show deficits in phonological processing, we know that most of dyslexia is driven by a dissonance between natural child's goals (i.e. comprehension) and adult goals (e.g. reading aloud to verify "correctness" of reading).
The long list of missing skills in phonological dyslexia will predominantly boil down to missing atomic memories such as letter:sound or sound:letter associations, sound blending, etc. All such atomic memories are susceptible to become toxic memory upon intense drilling or in conditions of stress. Even more so, the use of nonsense words in training or therapy provides no semantic reinforcement, which increases chances for toxic associations (e.g. see DIBELS assessment).
If the visual pattern, e.g. for letter Y, results in network instabilities on output, we may have a case of interference between sounds /waɪ/ and /juː/ (as explained in the toxic memory experiment in this text: Inducibility of dyslexia). A simple remedy is to build new associations, which usually requires forgetting of the wrong connections. In conditions of drilling at school or in a remedial program, repeated and frequent exposure results in a mis-consolidation of the unstable pattern. This troublesome association may generate stress on its own (without outside pressure). However, if it starts becoming noticeable by teachers or parents, it can result in a feedback loop that will tighten the instable memory consolidation and associate it with displeasure (see: decoding failure penalty). After a while, due to generalization, the mere input pattern that includes letter Y in any context may be associated with anxiety or fear of penalties. Reading become unpleasant or painful.
In the dual-route model of reading, phonological dyslexia stems from deficits in the phonological route (grapheme-to-phoneme conversion, visual → auditory route).
Surface dyslexia
Surface dyslexia is characterized by difficulty in recognizing words by sights (e.g. "yacht", "enough", "people", etc.). It is diagnosed in 10-15% of dyslexics. Most kids tend to focus on visual aspects of print due to the simple fact that print is processed by sight. This is why phonological dyslexia is a prevalent type. However, in their adaptation trajectory, some kids develop excellent phonological processing skills which may outweigh the visual capacity to process text. Paradoxically, such kids may develop an early interest in reading for its natural links with speech. In the course of learning, they may struggle with language irregularities and reject sight words. Those words may also get associated with the stress of reading difficulty. That's the first step towards surface dyslexia. Again, a perfectly healthy kids may be pushed or encouraged to undertake early reading and enter a positive feedback loop of dyslexia. Instead of improving fluency by reading for pleasure, the child may drill her weak sub-skills and make them worse by forming toxic memory.
In the same way as phonological dyslexia, surface dyslexia relies on a vast bank of reading memories. Those memories can be formed naturally while reading on demand and for pleasure. Any form of instruction may quickly lead to forming toxic memories instead. Those memories involve letter shapes, symmetry, sequences, sight vocabulary, etc. In some cases, phonological preferences may accelerate forming toxic memory if they are overridden by a teacher (e.g. when trying to first use phonics to decode or encode words, even if they are as simple as dog or red).
In the dual-route model of reading, surface dyslexia stems from deficits in the lexical route (direct visual → semantic route).
Deep dyslexia
Deep dyslexia occurs after a brain damage. It is characterized by semantic reading errors where a person reads one word as another word with a related meaning. For example, they might read "dog" as "cat" or "table" as "chair". Such substitutions may also occur in early stages of learning to read and are not a reason to worry. Deep dyslexia is rare and usually appears in people who previously had normal reading ability.
Comprehension deficit
When coercion leads to the first signs of dyslexia, heavy drilling in phonics may results in a kid who hates reading and automates reading. In phonetic languages, such a kid may read like a robot without a care about the meaning. This is a typical outcome of training at school under coercion where the goal of reading obscures the main goal that is learning, which rests on comprehension.
Type proliferation
The number of dyslexia types and classifications is overwhelming and growing. New authors continually coin new labels for old patterns. Misnomers abound, often leading to the inclusion of entirely different disorders, or even ordinary human quirks, under the dyslexia umbrella. In fact, I can invent a brand-new type right now: if someone statistically stumbles most often on the letter T, we can call it T-slexia. Dyslexia types proliferate along a number of deficits that provoke establishing new varieties of paid remedial programs. The following section is not too informative. It is illustrative with a dose of sarcasm.
Metaphor of limp types
Gallery of dyslexia types
Below is a sampling of some existing and semi-invented types, in case you're wondering whether you or your child might be suffering from one you just anxiously googled:
- Rapid Automatized Naming (RAN) Deficit
- Double-Deficit Dyslexia
- Visual Dyslexia
- Attentional Dyslexia
- Developmental Dyslexia
- Acquired Dyslexia
- Directional Dyslexia
- Primary Dyslexia
- Secondary Dyslexia
- Trauma Dyslexia
- Kindergarten Dyslexia. Yes, I made that one up but give it time.
If the trend continues, kindergarten dyslexia will be the next epidemic, diagnosed when a child fails to read Dostoevsky at age five.
Gallery of dyslexia deficits
Each newly defined subtype often stems from some observable "deficit" conveniently diagnosable by a sufficiently motivated specialist. Trouble might arise in:
- Naming letters
- Matching letters with sounds
- Recognizing rhymes
- Blending syllables
- Segmenting phonemes
- Manipulating sounds
- Noticing alliteration
- Achieving reading automaticity
- Decoding unfamiliar words
- Building vocabulary
- Spelling correctly
- Writing legibly
- Guessing strategically
- Persisting through difficulty
- Enjoying reading
- Acquiring background knowledge
- Learning foreign languages
Each one of these issues might be blessed with its own subtype, and naturally, its own custom-tailored (and billable) therapy.
Gallery of interventions
Naturally, with every new flavor of dyslexia comes a matching intervention: customized, trademarked, and preferably bundled into a subscription model. Where there's a new subtype, there's a new specialist, a new training course, a new certification, and a new therapy program, often accompanied by a research paper or a TED Talk to lend it legitimacy. This phenomenon may be described as intervention inflation, where the proliferation of therapies parallels the multiplication of dyslexia classifications.
This is the age of therapeutic personalization, where even a child’s confusion between b and d can be addressed through a 12-week neuroplasticity bootcamp. Results not guaranteed, but optimism highly encouraged.
Some examples of commonly marketed interventions include:
- Multi-sensory phonics therapy: Because regular phonics are not enough. Now children trace letters in sand while listening to Beethoven.
- RAN reconditioning drills: A treadmill for the brain's naming speed. Stopwatch included.
- Color overlay therapy: Tinted plastic sheets placed over text to allegedly correct visual processing.
- Executive function coaching: Also known as paying someone to remind your child to do their homework.
- Brain balance centers: Based on the premise that poor inter-hemispheric communication causes learning issues.
- Auditory integration therapy: For those who suspect reading issues originate from misunderstood musical tones.
- Forest-based literacy immersion: Reading barefoot in nature. I love jogging barefoot via forests. Without a book though.
Additionally, there has been a surge in digital solutions, including mobile applications that claim to enhance reading fluency through gamified phonics, eye-tracking, or AI-based progress analysis. Many of these tools emphasize user engagement, sometimes over actual efficacy.
Underlying all of these interventions is a critical question: Does the child truly require fixing, or do they simply need time, freedom, and a reduction in external pressure? In many cases, the act of labeling and intervening may obscure the possibility that the so-called deficits are natural variations in adaptive developmental trajectory, exacerbated by coercion in learning.
Misnomers
Let us not forget math dyslexia, a popular misnomer for dyscalculia. It earns a mention here not because it's a type of dyslexia, but because its association underscores how terminology spreads through cultural osmosis. In truth, both dyslexia and dyscalculia may stem from similar roots: attempts to learn under stress, coercion, and in conflict with natural curiosity. When the brain is forced to decode symbols before it grasps the meaning behind them, or before it sees the point, confusion is not a bug but a feature.
Conclusion
We are witnessing the inflation of dyslexia as a label, fueled by a mix of good intentions, diagnostic enthusiasm, greed, and a tendency to medicalize the natural variability of human development.
Comorbidities
There is little surprise that dyslexia often co-exists with dyscalculia or dysgraphia. Those conditions are also easily explained by toxic memory and would co-emerge in coercive schooling. Dyspraxia may be more complex, but its pop name "clumsy child syndrome" indicates that it may also emerge in conditions of stress when a child struggles with lacing his shoes or is ridiculed by his performance during a PE class.
Bangor Dyslexia Diagnostic Test provides another indication of the association between dyslexia and stressful environments. It includes classic toxic memory tests such as multiplication tables, the sequence of months, and even left-right discrimination.
Interestingly, ADHD is also on the list of comorbidities. ADHD is not related to toxic memory, but ADHD may also emerge in coercive education as a side effect of a creative mind (see: Confusing creativity with ADHD). No wonder then than compulsory schooling is a breeding ground for a whole host of dyslexia comorbidities.
Connectionist models
Researchers who study neural networks believe that these systems can approximate any function, given a sufficient number of parameters (weights and biases). This perspective naturally leads them to view the brain as a vast neural network. Their reasoning appears to be reinforced by recent advances in artificial intelligence, which can now mimic, or even surpass, humans.
However, there is a fundamental issue with treating the brain purely as a neural network. This issue highlights a key difference between biological intelligence and artificial intelligence and may help explain why AI remains computationally inefficient compared to the human brain.
Concept network model
The answer to this inefficiency problem lies in the conceptualization process, where portions of the network crystallize into specialized "processors". These processors become highly tuned to detecting and conveying specific patterns such as one’s grandmother. This idea is at the core of the grandmother cell theory, which is a not-so-popular perspective on how the brain processes information. Immature networks gradually develop semantic assemblies that evolve into a sparse localist agglomerates with the hypothetical grandmother cell being the ultimate product of the process.
Weak models
Behaviorists view dyslexia as a deficit in a collection of granular skills. Neuroimaging experts see the reading network as a set of specialized brain areas dedicated to phonological, grammatical, and word-form processing. From this perspective, dyslexia is framed as a problem in the wiring between these areas.
This approach can lead researchers to overemphasize phonological deficits or, at best, adopt the dual-route model, which accounts for both phonological and surface dyslexia. However, a more accurate interpretation of how the brain learns to read recognizes that these subnetworks develop dynamically and interactively. Their connections form and strengthen through distributed processing and feedback loops, rather than through rigid, predefined pathways.
This aligns with the connectionist model developed by McClelland and Rumelhart (1981 and earlier), which describes how the brain acquires language through emergent patterns of neural activation (see: example). In contrast, Pinker and Chomsky propose that language follows a set of explicit rules that must either be memorized or are innate. The connectionist approach challenges this notion, emphasizing learning as an adaptive, experience-driven process.
Conceptualist model
This ongoing debate mirrors the larger conflict between connectionists and proponents of localist representations or sparse coding. Similarly, we see tensions between the dual-route and connectionist models of dyslexia, as well as the long-standing "reading wars".
In the reading wars, sight word approaches align with the lexical route, emphasizing whole-word recognition, while phonics instruction relies on the phonological route, teaching readers to decode words through sound. However, a connectionist perspective would suggest that this division is unnecessary. Instead of rigidly adhering to one approach, the brain dynamically chooses the optimal route based on the context and the reader’s experience.
Viewing the brain as a concept network helps resolve these disputes. While any subnetwork may initially function as a vanilla neural network, over time it organizes into sparse concept maps and develops concept cells specialized in detecting a specific meaning in a pattern.
When the entire reading network is understood as a concept network, it becomes clear that we cannot precisely predict which pathways will develop their functions earlier and which will mature later. This process depends on individual learning experiences, shaped by both the brain’s current state and the nature of incoming signals from the environment.
In the case of dyslexia, the conceptual model suggests that a toxic memory can form in any link within the reading network. The severity of dyslexia will depend on the nature of the deficit, the number of toxic memories, and the degree of coercion that reinforces the pathological status quo, preventing the network from reorganizing and re-adapting.
Metaphor for model proliferation
Disclosure
You will find lots of materials about the types of dyslexia on the net. Look around for authors or links at the top or at the bottom. These are mostly commercial entities offering remedial programs. As I am part of a commercial company selling SuperMemo, let me then insist Do not use SuperMemo to remedy dyslexia!. Intuition is always the same: work on weaknesses. But that intuition is wrong. With SuperMemo you are likely to do more damage than good! I hope now you know that this is not a commercial message trying to sell you something.
Further reading
- Don't teach your child to read
- Educational dyslexia: most of dyslexia observed in coercive schooling is a result of coercion
- Dyslexia conspiracy: why my texts might give me a label of a conspiracy theorist
- Reading wars are over: Whole language vs. Phonics: children settle the reading wars with free learning
- Toxic memory: stress at school can result in dyslexia, dyscalculia, stuttering, and more
- Semantic learning and Asemantic learning: semantic brains reject knowledge of no measurable projected use
- Asemantic curriculum: curriculum that makes no sense to a kid will not be retained in memory
- Problems with special-needs education: teacher lobby generates special needs which generate special needs lobby
- Precocity paradox: late to talk, late to read? you may have a genius at hand!
- Optimality of the learn drive: why we can be sure that the child is always right in her educational choices
- Brain is a perfectly adaptable device: why it is the brain, not the teacher who is right
- Childhood amnesia: why early instruction in declarative knowledge is a waste of time
- Tunnel vision of school letteracy: in the era of YouTube, pushing phonics is doubly discouraging
- Accuracy in reading is costly: school perfectionism takes away all the fun of reading
- 100 bad habits learned at school: learning in a coercive school result in forming an alien human being
- Schools contribute to the dyslexia epidemic
Comments