Is speed reading really possible?

At around 7 or 8 years old, my reading level was several grades ahead of schedule. I was consuming large (for my age) novels with no sweat.

My parents noticed this and gave me a speed reading book. I incorporated each an every item on that book including the removal of subvocalization. It certainly gave returns right away, allowing me to consume books at a faster rate, but a huge cost.

Years later, I entered college as an engineering student. I noticed that when reading highly technical textbooks, and classes that required deep comprehension of text, I was much much slower that my classmates, or my comprehension level was much lower. Turns out, it was taking almost double the time for me to get through a section.

Really wished I didnt read the speed reading book now. I dont blame my parents at all, it was unfortunate that at nearly 40, i'm still trying to unlearn what that speed reading book taught me. I'm trying to re-learn subvocalication but its very very hard to 'slow' myself down. I've actually installed screen readers to replace subvocalization with some positive effect.

> You can't read without subvocalization.

While this article seems well researched, I find this claim very hard to believe. I can read while singing (not a song with words, but a tune that I know well, without words).

I really don't think I "talk to [myself] so quietly, it cannot be heard" and while my "tongue and vocal chords" may "receive speech signals from [my] brain", I'm pretty sure they're not executed and neither my tongue or my vocal chords move in any way.

(While reading in English there are some words I don't even know how to pronounce, so how could I "subvocalize" them...?)

- - -

The other point I have issues with is the fact that one should retain everything they read while speed reading. While I'm unfamiliar with speed reading classes' claims, I would argue that it's useful to be able to skim a big wall of text to

1- get a rough idea of what it's about

2- determine if it's worth a further, more thorough examination

That may not be called "speed reading", but it's certainly a helpful skill.

I took a speed-reading course and read War and Peace in twenty minutes. It involves Russia.

-Woody Allen

"I took a speed-reading course and read War and Peace in twenty minutes. It involves Russia." Woody Allen

The comments in this thread, specifically the questions about deaf readers prompted me to Google "deaf readers subvocalization", which lead me to the book "The Psychology of Reading" by Keith Rayner and Alexander Pollatsek.

I'll excerpt a few paragraphs.

[Page 191 - http://goo.gl/VLaaM]

Electromyographic Recording (EMG) records have been widely used to study the role of subvocalization in reading. [...] Generally, EMGs are also recorded from some other part of the body, such as the forearm, where muscle activity should have no relevance for reading. In fact, normal skilled readers show little forearm activity during reading, while EMG activity in the speech tract increases markedly during reading in comparison to the baseline condition (where the subject sits quietly). In contrast, deaf readers show a considerable amount of forearm activity during reading (we shall see the reason for that later in this chapter).

[Page 211 - http://goo.gl/R3NPT]

That profoundly deaf people can read at all is sometimes taken as evidence that speech recoding in normal readers is optional. However, as we have seen, they do not read very well, and the best available evidence indicates that when reading English text deaf readers recode the printed information into their native language (ASL) for comprehension purposes. Since they have not experienced speech sounds, recoding into a phonological code is precluded from deaf readers. Instead, the information is recoded into a manual form to aid comprehension. The poor reading of deaf people is thus probably due in part to inner speech being a more efficient system of recoding than overt manual gestures.

Readers of logographic systems such as Chinese can probably access the meaning of many printed characters directly from the visual representation. However, associations between the printed word and the appropriate pronunciation are activated during reading and appear to be important in comprehending text.

In so far as I have had "success" with speed reading it has been to force myself to run over the words more quickly and so interpret the commonly seen structures as single units. This means that if a piece is written in a very formulaic style then I can get through it fast because most of the stock sentences read like single words with only the unique modifiers jumping out of the page.

Think about when you look at code and automatically chunk the bits you have seen a million times before but are somehow magically able pick out the needle in the haystack which tells you what is special in this instance. The layout of code helps enormously for us, but I find the effect is the same with reading if I go quickly and with specific purpose to reformulate the information as it is scanned. And the more you do it the better you get because you have more data to draw from when identifying structures. It gets to the point where often I want an author to be formulaic rather than stylish and idiosyncratic because it makes it much quicker to reorganise and internalise.

Obviously, you get your best speeds when you're just accepting the text uncriticially. If you stop to think about it, you won't hit top speeds.

Beyond that, you can certainly train for optimal eye movement, local or global optimum as the case may be. After I took speed-reading, I knew not to look directly at the very ends of lines, and I found it more comfortable to center my eyes on white space than on actual text.

IIRC, I tested up to 800 WPM, but that was on grade-appropriate material in a high school I found easy.

(Of course, this was all in the 1970s, when what one read was black text in a familiar font in neat rows on white paper, all at least somewhat thoughtfully arranged. Things are different now ...)

My best real-life estimates of my reading speed was in the 400 words a minute range, with sub-vocalization. E.g., I have a few times read 3 novels the same day, most memorably 2 days before my PhD qualifying exams. (One day beforehand, I crammed like mad. :D One hour beforehand, I led my fellow students in the most awkward game of frisbee ever.)

But it all depends on the material, and the approach to reading it. E.g., there are certain chapter subsections of Hormander's book on Analysis of Several Complex Variables that famously take a week or more each to "read".

And yes -- I subvocalize.

Yes, it is possible, I do it, but not as good as I would like. You could read a book in twenty minutes as if it was a video with better comprehension that reading it slowly(as most of the brain disconnects with such a low bandwidth).

When you speed read you train your eyes to look over a wide area and you are able to read several lines at the same time and train the brain to make sense of it(use the subconscious to store-order the lines automatically).

So the best thing you could do is to control the formatting, so text is always the same size and always organized the same way, so if you force your brain to make sense of it, it does after time.

The best structure is newspaper-like text, lots of columns, small lines so your brain could make sense of entire sentences with just an eye shot.

It does not work with Hacker News, or most of the web by the way as each format is different and content is scrolled, with wide lines, not paginated.

It works really well when you could control the formatting, aka non DRM text you could interpret with your own software, but if you try to sell this software you will have problems when even blind people are not permitted to modify copyrighted text.

Wow. That was well worth a read, at whatever reading speed gives you full comprehension. The author did a good job of reading the previous scientific literature on the subject, and relating important issues to one another. He examined both historical claims (supposed world records of reading speed) and scientific claims (statements about how people read in general).

I read quite a few books about speed-reading when I was a university student in the early 1970s, putting the techniques to the test while taking courses in linguistics, foreign languages, history of technology, and Japanese literature in English translation. There are a number of good books about how to improve reading skills, with various levels of credulity about "speed-reading" claims. After my own research and experience, I have to agree with the paragraphs in the article submitted here based on more recent research:

"Ronald Carver, author of the 1990 book The Causes of High and Low Reading Achievement, is one researcher who has done extensive testing of readers and reading speed, and thoroughly examined the various speed reading techniques and the actual improvement likely to be gained. One notable test he did pitted four groups of the fastest readers he could find against each other. The groups consisted of champion speed readers, fast college readers, successful professionals whose jobs required a lot of reading, and students who had scored highest on speed reading tests. Carver found that of his superstars, none could read faster than 600 words per minute with more than 75% retention of information.

"Keith Rayner is a professor at the University of Massachusetts, Amherst and has studied this for a long time too. In fact, one of his papers is titled 'Eye movements in reading and information processing: 20 years of research,' and he published that in 1993. Rayner has found that 95% of college level readers test between 200 and 400 words per minute, with the average right around 300. Very few people can read faster than 400 words per minute, and any gain would likely come with an unacceptable loss of comprehension."

I figure that my comfortable, steady reading speed in materials on a wide variety of subjects at an upper-division undergraduate to graduate school level is about 500 words per minute, with good comprehension. I plainly don't need "tl;dr" summaries of articles submitted to HN as often as many HN participants ask for those. (To be sure, many HN participants read English as a second language, and we should admire people who come here to participate in a second language, something very few Americans could do in a non-English-language online community.) I definitely read more slowly and with less comprehension on the first pass in Chinese or in German, my two strongest languages for second-language reading, but I have read whole books in both of those languages for fun or for research. I have diminishing ability to read other languages that are mentioned in my user profile here.

tl;dr: Don't worry about fancy eye movements too much, and don't worry about subvocalization too much. Just read steadily and think about what you are reading while you are reading it for best memory of what you read and best comprehension of what the author was trying to say. Building up your vocabulary--by more reading, of course--is the best way to build up your reading speed.

AFTER EDIT: The claim in another subthread here is specifically WRONG that Chinese constitutes any kind of proof about subvocalization. I'm not committing to a position on whether or not subvocalization, as defined by throat muscle movements, always occurs in reading, but I know from the books Visible Speech by John DeFrancis (a scholar of Chinese)

http://www.amazon.com/Visible-Speech-Asian-Interactions-Comp...

and Reading in the Brain: The Science and Evolution of a Human Invention by Stanislas Dehaene (a neuroscientist who does brain imaging studies)

http://www.amazon.com/Reading-Brain-Science-Evolution-Invent...

and from my own study of four different Sinitic modern languages (Mandarin, Cantonese, Taiwanese, and Hakka) that all writing systems, properly so called, are systems for writing out speech. Writing is based on speech everywhere in the world and the Chinese writing system is full of clues that most written characters are based on the SOUND of spoken morphemes.

How you might write the conversation

"Does he know how to speak Mandarin?

"No, he doesn't."

他會說普通話嗎？

他不會。

in Modern Standard Chinese characters contrasts with how you would write

"Does he know how to speak Cantonese?

"No, he doesn't."

佢識唔識講廣東話？

佢唔識。

in the Chinese characters used to write Cantonese. As will readily appear even to readers who don't know Chinese characters, many more words than "Mandarin" and "Cantonese" differ between those sentences in Chinese characters.

I checked the link to The Straight Dope about the reading comprehension tests they had conducted; turns out they hadn't conducted one at all. So I guess the author didn't read slow enough before writing.

I've discovered a few things about reading speed in my experience.

1) When I read fiction by good authors, I take great delight in reading their sentences, so I slow down drastically to take every word and piece of punctuation in.

2) Reading speed improves with practice.

3) Read smarter: if you don't need 100% comprehension then read topic sentences of paragraphs carefully then skim the rest of the paragraph for details you think are useful. And in news stories, they follow the "inverted pyramid" for paragraphs meaning that they're ordered such that the first paragraph is the most important, second one is second most important, etc. So you can stop reading at any time.

4) Choose a column width that suits you. A lot of time is spent on line transitions. Smaller columns makes this easier which you can control via your window size.

I read this at 800 wpm (adjust this down @ 10-20% if a test I did a while back still stands) fairly comfortably using http://www.spreeder.com/app.php. I find it very much easier to stay "in the zone" using it.

Speaking of speed reading, the current layout of HN is not optimal for reading. Main content width is way too long which makes of long sentences. Font size is too small - so you're forced to zoom in.

Perhaps PG could think about improving readability of the site.

I did a speed reading course when I started uni and yes it does work. A lot of stuff is bad habits - some others have mentioned, sub vocalisation, keeping your eye in the middle of the page and read with your peripheral vision, obvious stuff - don't slide your finger along the page, don't use a ruler (some do) and so on. I improved my reading and comprehension quite a lot, but find I turn it off when reading for pleasure versus reading to extract facts. It's well worth doing - even if it only increases your reading speed 10 or 20%. [edit: How to skim read was another useful tool]

I did a speed reading course in high school and often wondered if it was a scam.

Not a scam, scam more a motivational speaker/Eat less exercise more style scam. If you followed it through it would work but almost no one follows through hence the course/advice becomes bunk.

For instance one thing we were taught was to first skim the book/chapter really quick then read it 'properly'

Makes sense to me, but I never do it.

People do have different reading speed levels and I don't think it's all innate so I don't think it a crap idea.

I would agree that maybe the teaching is currently more based around customers and less about pedagogy though.

When I was younger I took a reading test and scored just over 1,000 words per minute with 100% retention.

I remember reading an article that there were 3 types of readers.

Readers who sub-vocalize the words

Readers who "hear" the words

Readers who "see" words

Not sure if that's an accurate assessment, but I definitely don't sub-vocalize. If I sub-vocalized, I think would at least some general idea how to pronounce unusual character names from fantasy novels. I couldn't have even tried to pronounce most of the character names in Lord of the Rings, immediately after reading it but I could recognize the name patterns if I saw them.

I think speed reading is most useful when you are reading some text where only parts are relevant or new to you. By quickly reading those, you can focus on the important pieces and spend more time to think about them.

Interestingly, the faster I read, the less I'm able to "filter" the information I read about, it seems almost as if my brain absorbs it directly. That means that I'd never speed read content that might contain bad memes or false information: political propaganda, religious texts, etc.

Does anyone know a good test for reading speed (including comprehension)?

I've been a bookworm from early age and even though I slow down when the material is highly technical and requires better comprehension I don't find the words being repeated in my head. As a matter of fact, while reading I find that while I do talk in my head, I might summarize or ask myself questions about a paragraph (stuff like: "Don't believe. Rem. check refs", like I did a few times in this article), as opposed to repeating the words I'm reading. The only moment I find I'm repeating words mentally or sub-vocalizing is when I'm writing, and I believe I do so to get the underlying tone of what I'm writing.

When I read a novel, I tend to want to spend more time immersing in the world and the events in it. While I slow myself down in this case, I tend to make images in my head while reading and I've found that it becomes inherently harder to do if I'm subvocalizing... it removes focus from the imagination side of things for me. I don't really believe that it's impossible to read without subvocalization.

In any case, one of the tricks I use to speed read at a moderate speed (as opposed to skimming) is to black out mentally everything but the current line and process it as a whole. It works like imagining a rectangle enclosing the current line. Works wonders for me taking into account I'm the fastest reading person in my social circles.

there are plenty of books around this topic. In short there are a at least 4 ways speed reading that you use for different materials. As the other comment pointed out the 1st thing you have to do is stop vocalizing what you read, then you need to learn vocabulary so you would know all of the words without putting much thought to it.

speed reading can really be a game changer, I used to get bored of books myself, but when I can read faster the experience seem to be way more enjoyable!

I worry more about the quality of my comprehension than the speed at which I read. I"d rather take ages and get a good understanding than speed up the pace at which I read. Depth over breadth any time. In product features, in overall understanding, in capability, in testing, art, in all things life. This is akin to a get rich quick scheme, do things on the cheap, doing more with less. You can only do less with less.

I'd argue with a few points made in this article. Subvocalization can most definitely be minimized - Not removed completely, but minimized. The article brought out a study from NASA stating that there were minute muscle movements, which I would believe - but if you can minimize it to a dull effect, can drastically improve your speed of reading.

As with "fancy eye movements" as some have claimed it - Whether they have much of a help, I'm not sure. Personally I generally read with my right eye, and my left eye tails a fraction of a second behind which generally helps pick up a few details I may have missed - But I notice I have trouble with this if a page's width is too large (HN has a fairly large width and I will read much faster if I make the window smaller). Anything more than that seems as if it would be fairly difficult - But I haven't read into the various techniques.

Can you learn to read at the crazy speeds mentioned in the article? Probably not. Speeding your reading up though seems definitely plausible - At least it seems to work for me, but I can't draw a very good conclusion from just myself in the sample.

I used the coursebook Breakthrough Speedreading successfully in teaching myself the technique a number of years ago. Though all arguments here about nuance and critical thinking also were nagging in the back of my mind then, the actual reason that made me drop it was actually reading another book on an unusual technique: Relearning to See on the Bates method.

The Bates Method emphasizes that one of the worst offenders to natural vision is staring, or the act of fixing a point and relying on peripheral vision to capture a much bigger area. Relearning to See is in big part about relearning to skip, like animals do, when they constantly move their eyes and heads to scan the environment or an object. And speedreading was at odds with this, because it emphasized you should fix your vision on some center point in the page and allow your peripheral vision to gobble up the surrounding text, and you need not even do it sequentially, you should get to the point where you "understand" the image of the page as a whole. That's the bill, anyway. So I opted for natural vision (and comprehension).

As I was reading this article, I used the SpeakIt! Chrome Extension[1] with the native voice setting to gauge how fast the quoted speeds actually are.

For me, 800 wpm was about the limit of how fast I could read while still understanding most of it. That's the '4x' setting. The '2x' setting is 400 wpm, which based on the article and my experience is a good target reading speed.

It's not entirely perfect because the rhythm imposed by the text-to-speech engine is not optimal for fast reading – it pauses when you don't need a pause and powers through complex sections, but I found it an interesting exercise nonetheless.

Lately I do most of my reading by listening to audiobooks. The last two books I read, I experimented with the 2x speed setting on the iPhone. It took a few minutes to get used to, but now I can't recommend it highly enough. Now I can get through audiobooks in half the time and I've totally adjusted to the butchered sound quality.

[1]: https://chrome.google.com/webstore/detail/speakit/pgeolalili...

Anecdotal personal stuff: I learned to stop subvocalizing in grad school. I had no idea it was a canonical speed reading strategy. I can read really fast now, with unimpeded comprehension.

Common sense stuff: The author says that scientists say that reading without subvocalization is impossible. The fact that deaf people can read is extremely inconvenient for this theory. Seriously, man?

It's funny this comes up here and now, as I am working through "How to Read a Book", which was in one of the recommended reading threads on HN recently. In HtRaB, they talk a little about speed reading, but dismiss it as mainly something that can help bring someone up to speed, ie, it is mainly helpful in a remedial sense. I think the commentary with comprehension is quite cogent (it's precisely why I'm reading HtRaB), and so therefore the discussion should turn to what to read, since time is limited. Again, I am grateful to previous threads on HN for pointing out good reads, and feel that a more considered approach to how time is spent (both in what is chosen to be done and how well it is done), rather than in how quickly something is done is most beneficial.

There's another Ronald Carver book called "Reading rate : a review of research and theory", which comes to the conclusion that speed reading doesn't work.

People can skim at faster rates than they can when reading with full comprehension, but then their comprehension will suffer. You can learn to skim, to take notes, to get quick overviews of whatever you're reading by noting things like chapter and section headings and by reading for the main ideas of sections and paragraphs.

These are all useful skills. But they are a far cry from being to read every single word in the text with full comprehension, and doing that at a significantly faster rate than normal. This latter achievement is something there is little or no evidence of.

The whole eye-moving thing is a complete waste of time and effort. The real trick to speed reading is RSVP[1], rapid serial visual presentation.

In RSVP the words or small groups of words are flashed in extremely fast succession without the need for the reader to move their eyes at all.

I use iRSVP on my iPhone to read books from my Calibre library and my reading speed has never been higher. It's not for everyone and requires a high degree of concentration, but it enables me to read and understand books much faster that I would have been able to otherwise.

[1] http://en.wikipedia.org/wiki/Rapid_serial_visual_presentatio...

Wondering about subvocalization, what about reading in a language I never learned to speak?

I agree with the conclusion that improving reading comprehension is the best way to improve reading speed, at least for me. There are two things I learned from my stuttered attempts at speed reading: (1) The best way to increase reading speed for me is to simply stop re-reading sentences that I already read. I do this unconsciously usually because I didn't fully understand the sentence the first time. Better reading comprehension reduces this. (2) Attempting to speed read is probably a waste of time for me because most of the material I read is technical, so comprehension is usually the bottleneck to getting through the material, anyway.

Different people have different minds/brains. My ex read so fast that trying speed reading techniques slowed him down. I read ridiculously slowly for someone who was in a lot of gifted classes. I tried speed reading techniques. I was able to read faster. I hated it and went back to my snails pace voluntarily. I read with good comprehension, but, yes, I "say" everything mentally.

My oldest son also reads shockingly fast. When I began watching him play videogames, he flipped through dialogue so fast I could not read it. I thought he was just skipping familiar parts. Nope. He just reads that fast. He has to slow down if he wants to include me.

I have 3 thoughts on speed reading

1. Speed reading fiction is about as sensible as watching movies in fast forward.

2. Speed reading non-fiction will not lead to much formation of conceptual knowledge in your brain, and definitely not the sort of rich inter-connected learning that we probably read for in the first place.

3. Speed reading where you develop a sixth sense for which parts of, say, an academic paper you should slow down and read carefully could probably be a good thing.

In general, while I used to think there was some sort of prestige attached to how much / how fast you could read, I don't think that anymore. It's how 'deeply' you read that counts.

I once heard a talk by an top government official who says she has to read a briefing book of hundreds of pages every single day from 10p.m. to maybe 1a.m. Of course, They skim a lot and they probably know a lot about the issues already but still they must read at a much higher speed than a college kid to finish that kind of reading assignment every day, right?

I know some people who read a book a day and I can't do it a week, and none of them need to advocate their reading speed. They must read a few times faster I do, I think.

Does anybody here read new texts, with no skimming and a decent comprehension at 1000wpm?

How do deaf (since birth) people subvocalize?

Around 5th-7th grade, I was in advanced reading and part of a speed reading training class/experiment. It still all seems a bit odd to me. A woman came in a couple times a week with a special projector that projected a book on the screen, but using a rapidly moving light, only showed a few words at a time on the screen. Each week the light moved faster. This was followed by comprehension tests.

I read about 30% faster than most people, but I have no idea if this is really why. Was anyone else a guinea pig? This was in Central NY in the 80's.

So, whenever I think of the subvocalization debate, I immediately see this video. The basic idea was, do not presume that everyone does the same thing you do in order to achieve the same goal.

https://www.youtube.com/watch?list=PL273BA8C6E484C988&fe...

In other words, I find it hard to believe a claim that subvocalization is required across the board. At least, at an absolute level.

After reading this very informative article, it occurred to me that developers of new tools that will be built upon heavily would do well to try to think of names that can be vocalized in one or two syllables. Reading the name PHP or the mouthful WYSIWYG should take longer than words such as C or Java or SCRUMM or perl if we are indeed vocalizing them. Perhaps the convention of adding www to the beginning of website names has had a substantial impact on the efficiency of web navigation?

It's really up to the text. When I read math or comuputer science papers, I have to read slowly, and a few times over, to get any comprehension, even from the non-formula prose.

I wonder if NASA has done comparative studies on native vs foreign language wrt sub-vocalization and reading/speed reading ?

It is rather well known that native language and (all) foreign languages are "stored" in separate parts of the brain (with some limited controversy over whether there exists "bilingual" people in a biological sense).

I wonder if that has an effect on how we read written text (other than that most people presumably read faster and with better comprehension in their native tongue).

Good speed reading courses and books not only teach you how to skim faster and not to subvocalize but how to comprehend and learn better. For example in Paul R. Scheele's book he teaches you how to start on a book:

read the table of contents,

then read the headlines through the book,

then speed read the book and then try to probe your comprehension, ask questions etc.

The table of contents and the probing is the most important. Teaches you recursive learning, when you first build a foundation then you fill out the details.

I also use spead reading once in a while. I can read realy fast going through large documents To be able to find something quickly I once got a huge document on my desk 3400 pages, and i had to decide if we would do that project, please try to read it. To their surprice i fineshed reading and pointed them on the hard parts of it, they never thought anyone would be able to read the entire doc in just a few hours for the deadline.

Around the point in the article describing the eye movements required for reading, I experienced a weird anxiety as I attempted to both perceive my own eye movements and consume more of the text. After a few sentences it actually became uncomfortable. Gave me the impression that implementing speed reading strategies would be dubious for me, personally.

I'm surprised about the bit about the subvocalization; I always thought that brain speed was the number one limiting factor.

1) Silence the inner voice.

2) Don't fallback with your eyes, just re-read the entire sentence.

I can combine two lines I've read into a sentence while reading the next two lines. It helps when I'm in an exam, or when looking for something in a Website.

I heard of people who scan pages diagonally, which makes my envious :)

Besides sub-vocalization I tend to read loudly making full use of my oral musculature. It helps me stay active, improves comprehension and retention. I am not sure if it's a bad habit or just the way my brain works. Any of you guys have same experience.

Just a shout out for forcefeed - a javascript bookmarklet that lets you read any web page without moving your eyes. I think it's incredibly helpful, even if my comprehension isn't 100%.

Yes, if you don't value what you read and just want to get a quick idea of it.

It's not possible for actually _reading_ stuff.

Which reminds me of the Woody Allen quote:

― "I speed-read "War and Peace". It's about Russia, right?".

Just a shout out for forcefeed - a javascript bookmarklet that lets you read any web page without moving your eyes. I think it's incredibly helpful.

Is there any good speed reading online course?

For a user of a speed reading CD (eye-Q), thank you for this info!

Not for me.

The text says it's been proven by NASA and several others that you can't remove subvocalization, yet most speed reader comments here go on to say "yes it works i removed subvocalization". It's obvious all these "speed readers" didn't read a single line of this article.