4 ms·

I think our brains just picks all the letters in a word, mixes them up together into something like "e:a:l:p = 1:1:1:2", adds context and does a lookup into a d

I think our brains just picks all the letters in a word, mixes them up together into something like "e:a:l:p = 1:1:1:2", adds context and does a lookup into a dictionary of such "letters mix" -> meaning pairs. If this is correct, dyslexia is when the brain can't quickly estimate the "amount" of each letter in a word, and the incorrect proportions result into an incorrect lookup.

slg 6y ago
That analogy is close but doesn't quite match my experience because there doesn't have to be much overlap in letters in a word. I went back and reread a few of my most recent HN comments looking for an example (my condition impacts both reading and writing). There are some easy to understand mistakes like replacing "went" with "want". However I also used "because" in place of "beyond" in one comment. Those two words start with the same two letters, but they are completely different after that and are relatively distinct visually. An algorithmic analogy I would use would be fuzzy matching. Most people's brains are set to look for exact matches (or fuzzy matching with a super high match criteria). When one part of their brain needs access to a word, the other part of their brain will supply the exact match. My brain is using fuzzy matching with the confidence level set mistakenly low. When part of my brain needs a word, the first match returned is occasionally a false positive match. The rest of my brain just accepts it and continues onto the next requested word. The match quality isn't a traditionally fuzzy matching algorithm like Levenshtein distance. It is heavily weighted to both the start of the word and the expectedness of the word in the given context. The latter makes it very difficult for me to immediately notice my own written mistakes since I am already primed to expect what I intended to write and not necessarily what I actually wrote.
- linuxftw 6y ago
  Your second paragraph seems quite insightful. I will occasionally write out the wrong word, only to re-read the incorrect word as the word I intended to write, totally missing my mistake. As you describe, I think most often it relates to the 'fuzzy matching' of the first few bits of a word, my brain auto-completes the action of selecting the wrong letters for the word as the intended word. I seem to be able to break the spell in a number of ways. When I'm re-reading a final draft of something (say an important email) I will try to eliminate context and deliberately read each word individually. Another way to break this spell is to read a document much later in time. Sometimes the pattern matching self-corrects later, or seemingly I have forgotten the verbatim phrasing of the idea that I previously wrote, and so upon reading I pick up the mistake. I read thousands of words per day, probably write well into the hundreds. Fortunately, the amount of errors for me is very low, but if this were to happen with increased frequency, reading and writing would be quite stressful.
  - slg 6y ago
    Yep, trying to give myself extra time to forget the exact intent of the writing and revisiting it later was my primary approach through school. However the absolute lifesaver since then has been text-to-speech software. Listening to my writing being read back to me makes it absolutely obvious when I used an incorrect word.
    - linuxftw 6y ago
      Interesting analysis. I think text to speech could help many people review their writings in this way. Sometimes things just have to 'sound' right.

slg 6y ago

That analogy is close but doesn't quite match my experience because there doesn't have to be much overlap in letters in a word. I went back and reread a few of my most recent HN comments looking for an example (my condition impacts both reading and writing). There are some easy to understand mistakes like replacing "went" with "want". However I also used "because" in place of "beyond" in one comment. Those two words start with the same two letters, but they are completely different after that and are relatively distinct visually. An algorithmic analogy I would use would be fuzzy matching. Most people's brains are set to look for exact matches (or fuzzy matching with a super high match criteria). When one part of their brain needs access to a word, the other part of their brain will supply the exact match. My brain is using fuzzy matching with the confidence level set mistakenly low. When part of my brain needs a word, the first match returned is occasionally a false positive match. The rest of my brain just accepts it and continues onto the next requested word. The match quality isn't a traditionally fuzzy matching algorithm like Levenshtein distance. It is heavily weighted to both the start of the word and the expectedness of the word in the given context. The latter makes it very difficult for me to immediately notice my own written mistakes since I am already primed to expect what I intended to write and not necessarily what I actually wrote.

linuxftw 6y ago
Your second paragraph seems quite insightful. I will occasionally write out the wrong word, only to re-read the incorrect word as the word I intended to write, totally missing my mistake. As you describe, I think most often it relates to the 'fuzzy matching' of the first few bits of a word, my brain auto-completes the action of selecting the wrong letters for the word as the intended word. I seem to be able to break the spell in a number of ways. When I'm re-reading a final draft of something (say an important email) I will try to eliminate context and deliberately read each word individually. Another way to break this spell is to read a document much later in time. Sometimes the pattern matching self-corrects later, or seemingly I have forgotten the verbatim phrasing of the idea that I previously wrote, and so upon reading I pick up the mistake. I read thousands of words per day, probably write well into the hundreds. Fortunately, the amount of errors for me is very low, but if this were to happen with increased frequency, reading and writing would be quite stressful.
- slg 6y ago
  Yep, trying to give myself extra time to forget the exact intent of the writing and revisiting it later was my primary approach through school. However the absolute lifesaver since then has been text-to-speech software. Listening to my writing being read back to me makes it absolutely obvious when I used an incorrect word.
  - linuxftw 6y ago
    Interesting analysis. I think text to speech could help many people review their writings in this way. Sometimes things just have to 'sound' right.