Slashdot Mirror
The Science of Word Recognition
neile writes "I stumbled across a fascinating paper over at the Microsoft Typography site today that provides a really nice overview of the different theories on how humans read. If you thought we read by recognizing word shapes, think again! With the assistance of fancy eye-tracking cameras researchers have been able to devise several clever experiments to give us new insight into how reading works." We've linked to some of Larson's work previously.
Would one of those stupid comments about the colour scheme on /. be on-topic now?
Wh47 d1d j00 541, 31337 15n't t3h r0xor5 ne m0r3???
I was reading what was written on her T-shirt!
So are Microsoft going to patent the way we read and then sue?
"If you are reading this then you owe Microsoft royalies"
----
To make it more obvious, stick a tall letter in a word that only has short letters and you'll come away thinking word shape does matter.
(or did he explain it... there were way to many words and way too few glossy pictures in that article for me to comprehend it)
New technology will soon be revealed that will instruct Slashdot users on the proper spelling of "lose".
The USSGN (Union of Slashdot Spelling and Grammar Nazis) is expected to stage protests against the new product in the interest of keeping their jobs.
It would be cool if it didn't suck.
With the assistance of fancy eye-tracking cameras researchers have been able to devise several clever experiments to give us new insight into how reading works."
Oh they must have been using EyeQ....
I can read at 44692 words per minute! Thanks for posting that long article for me to read, I needed the exercise.
And thank you EyeQ! Your the greatest!
Really though, they say that the more letters/words mean faster reading times. It's true. Think about a book or article you've read. When the words are together on the page it's easier to read because your eyes can jump around letting your brain fill in the blanks.
Ever read something that made sense but you couldn't quote it word for word? It's likely because you read in this same way.
Get your Unix fortune now!
"Evidence from the last 20 years of work in cognitive psychology indicates that we use the letters within a word to recognize a word."
Man, I'm so glad they finally figured this out...
Does anyone else think that merely analyzing how english is read is very closed minded? I'm pretty sure only a very small percentage of the world speaks and reads english.
I would love to see a study comparing how english is read to how chinese is read by native speakers. Very interesting i would gather.
A Fatal OE Exception has occurred, Sig will now reboot.
While reading the article, I suddenly become hyper-aware about how I was reading the article. :-)
Don't let the Microsoft name scare you off - the article makes for a fascinating look (pun intended) into how we read. I wonder, though, if these findings are duplicated with written Oriental languages.
You call this a signature?
Since most people in the world don't use the latin alphabet, it would be interesting to find out how word recognition works for them. And how they read words in our alphabet.
***Quis custodiet ipsos custodes***
The final conclusions are similar to what I learned in my college linguistics classes 15 years ago. Language contains a lot of redundancy. The reason is that we often encounter situations of so-called "reduced redundancy". For example, someone might have sloppy handwriting so you can't make out all of the letters. Or you might be talking to someone while they brush their teeth. If language were highly optimized, we wouldn't understand a thing in these situations, but because of redundancy we can usually communicate very effectively.
The same applies to reading. The conclusions of the paper seem trivial to me. Of course, reading exploits "visual" and "contextual" information. How else would be understand a sentence like "The boy ate a ham___er" (with a few letters obscured)?
The fact that the brain's neural net adds up the weighted lexicographic, syntactic, semantic (and even pragmatic) information available to it in order to interpret language should be familiar to anyone who's read Goedel, Escher, Bach. And that was published in 1979...
Peer Pressure
Hree is an epamxle of jsut taht, it's qitue esay to raed, ins't it? Agulohth it can get plluartraicy hrad wtih the lgnoer wdros.
-- Intelligence is soluble in alcohol
when are they going to repeat these experiments in let say China or Japan? I'm *very* interested in what would the conclusions be there. ...
For what i know abaout japanese, they don't use spaces between 'words'. A single kanji represents the whole word and their outline is always more or less square. So the whole bouma theory fails here, as he finds out.
I'm sure they could leard more interesting things in other writing sysmtems
Research shows that
I found myself becoming aware of how I read while I read. Fun! I agree with the author regarding letter recognition. The parallel aspect of word recognition is very interesting as well because it begins to explain why we are albe ot raed srcambled txet os eaisly!
Also, more work needs to be done to consider the visual cues outside the focus of attention. It is here that, I believe, shape and form cue the reader, more than letter shapes do, as to the potential content of the text to come. (Exactly how is for the geniuses.)
Blogging because I can...
Read it, it's interesting. It does get a wee bit weird when it's describing how you read as you read... a sort of super-conciousness about my eye movements. It's like when you become aware of your breathing or something and then have to conciously pay attention to it for a while to make sure it doesn't stop!
-- *~()____) This message will self-destruct in 5 seconds...
While some of the results here are interesting (but old), the fact that the entire study focuses on exactly 1 script and 1 language basically renders the conclusions worthless (as conclusions about cognition in general... I suppose they still have value as conclusions about English and the Latin script).
What has happened here is:
1 -- Observe people reading a given language/script
2 -- See how they make use of features of that particular language/script, such as tall letters, case, and the occurrence of 'skippable' words such as articles
3 -- Describe the way they use these local features, and call that a theory of reading in general.
I don't really understand how to apply a theory of reading based on word and letter shapes when there are so many people reading text in which:
--There are no letter boundaries, and/or
--There are no word boundaries, and/or
--Letters all have the same form factor
The experiments described would probably generalize very well to arabic and greek scripts, pretty well to cyrillic (no tall/short letters to speak of), badly to devanagari-type scripts, very badly to Chinese and Japanese, and not at all to hieroglyphics (though I agree that there may never have been a reader of hieroglyphics who was fluent by modern standards).
To pretend that these experiments apply to humanity in general rather than the author's own language/script choice is silly. It's an interesting article and I'm glad the research was done but unfortunately a certain failure to 'get' the multilingual nature of humanity, which I don't really expect to find in MS work, is in evidence here.
Whence? Hence. Whither? Thither.
Although it is nice to see mentioning of my trade a /., this paper has about the status of a student's essay. It doesn't even mention literature after 1998!
http://news.bbc.co.uk/1/hi/health/3618060.stm is a good read about dyslexia it isn't exactly related but you might be interested :)
>renerding on firefox
re-nerding! ha ha. Best... typo... ever...
They will never know the simple pleasure of a monkey knife fight
If there's one real take-home lesson of brain-design from cognitive science, it's that the brain tends to do everything several different ways in parallel, and then use the results from all of them.
Obviously it can't all be shape, there are plenty of words with identical shapes and yet these are distinguishable.
But it could certainly be true that we use shape and parallel letter recognition at the same time. Shape narrows the field of possibilities from millions to a small handful, and then parallel recognition chooses one of the options.
Whatever happens, you can be sure it's terribly complicated, extremely robust and very efficient.
From the article: ...lowercase text is read faster than uppercase text.
This could also explain why nobody likes to read email where the other person uses all caps.
dunno, firefox / moz has one of my favourite features
tools
great for annoying "web site designers" who can't design for shit
There are places where the networks are not touching,and there are places where they are-Boeing's Lori Gunter
I'm surprised this guy is actually working with ClearType. That is just a simple way of making characters appear better by using sub-pixels to increase character resolution. I would think this type of work would be better applied in optical character recognition, maybe even with cursive handwriting.
The FArticle does, in fact, address this, though not directly - it puts forth a theory that all letters in a word are absorbed simultaneously, and the brain re-orders them. This is given as theory #3, admittedly a ways down.
This gets me thinking, though, about the importance of context. If you drew the letters PLEORBM in a Scrabble game, it might take a while to see the word staring at you. But in the context of a (mangled) sentence: "you can sitll raed tish wouthit a pleorbm," it much more easily jumps out. Interesting.
Infants of English-speaking parents easily grasp the Korean distinction between a cylinder fitting loosely or tightly into a container. In other words, children come into the world with the ability to describe what's on their young minds in English, Korean, or any other language. But differences in niceties of thought not reflected in a language go unspoken when they get older.
Absolutely. And adults can "relearn" those distinctions, too; I found that as my Japanese studies progressed (started at 19, pretty close to native now) the range of things I was able to think about expanded considerably--so much so that now I sometimes have trouble speaking to people in English because English doesn't have a word for the concept I'm thinking about.
If there's those that have shied away from Microsoft, well because they're Microsoft, you might not be aware of http://research.microsoft.com which regardless of which side of various fences you might sit has some very interesting material and is generally worth tracking over time.
Aplogise for the tangent, on the back of this article seemed an apt place to point to the MS research site for those that might not of been aware of it.
The internal representations for these models convert the letter information to phonemic information, which is seen as a mandatory step for word recognition. It is well known that words that have a consistent spelling to sound correspondence such as mint, tint, and hint are recognized faster than words that have an inconsistent spelling to sound correspondence such as pint
I can not believe this is in a serious paper. Mandatory? Please. What about people born deaf? Are they all unable to read?
I'm wondering about the competence level of the readers used in the various tests. People have, in my experience, a wide range of reading ability levels ranging from those who still have to "spell out" many words letter by letter up to speed readers who can read entire phrases (or even whole sentences) as easily as most people read single words. If we divide them into three groups (phonetic readers, whole word readers, and "cognitive chunkers"), would these results be consistent from group to group?
:)
Learning to read, like learning higher math, is a process of internalizing certain reflexes. Most people alive today do not understand calculus. Most will also never learn to read much faster than they can speak aloud.
Ultimately, 80% of the people voting in the next presidential election are of average or lower intelligence.
The rest of us are Slashdot members.
TLR
A man no more knows his destiny than a tea leaf knows the history of the East India Company
As a non-native (but fluent) speaker of English, and the husband of a fluent English speaker learning Danish, I can tell you quite well that there are many concepts that have a single word describing them in Danish but not in English, and vice-versa. Some words are normally considered equivalent but have slightly different extents ("pink" covers more colors than the common translation "lyserød", for instance).
The grandparent also didn't say "couldn't be expressed", but "has no word". Given enough verbiage, you can (probably) express any word in one language in any other language, but that's not what you want to do in conversation.
And if the "language of Shakespeare" is so all-encompassing, why has English since then been stealing words from other languages like a slum rat during a riot in a shopping mall? Mind you, I think this is a good feature that adds expressiveness to the language, but it clearly shows that there are things that English speakers consider important enough to be able to express succinctly that they'll bring in foreign words for it.
-Lars
---
While the study is certainly about reading English texts, could one draw some conclusions about the readability of source code? I guess at least the finding that whitespace governs the jumps of our eyes might have some relevance here.
The Tao of math: The numbers you can count are not the real numbers.
I am somewhat fluent in Chinese. Though syllables in Chinese (and Korean) approximately fit into squares, they share two characteristics with alphabetic word shapes:
First, Chinese characters are often composed of several smaller characters, 500 or so, instead of the 70'ish letters and numerals (including capitals) in English. We say such a character may have a "moon" sub-character on the left, a "white" on the right and so on. The sub-characters can be partial clues to meaning and pronunciation (e.g. a left moon usually signifies part of a body, and the a right moon means its sounds like "bai"), but there are no steadfast rules. Just like in English where the pronunciation can vary from the spelling, and the whole meaning vary from those of the prefix and root. But Chinese breaks the rules more often, probably since many of the characters have been around 3000 years- seven times longer than modern English spelling rules. The Korean writing system is totally planned and recent, so it is very logical. The sub-characters represent the beginning, vowel, and end of a syllable, gracefully packed into a square.
Second, you can trace the boundaries of Chinese characters too and see distinguishing characteristics. They might have a gap in one corner, a ragged stroke defining an edge, etc. Just like in English words a learner will move from observing the strokes and sub-characters into the gestalt of the whole character. Just like any other language, chinese characters are contextual. Combinations are sematically constrained to one to four syllable word. And they are grammatically constrained to expect nouns, verbs, modifiers, etc, in certain sequences.
I am sure that we've seen this e-mail floating around. Doesn't it seem like we read in shapes?
I cdnuolt blveiee taht I cluod aulaclty uesdnatnrd waht I was rdgnieg The phaonmneal pweor of the hmuan mnid Aoccdrnig to a rscheearch at Cmabrigde Uinervtisy, it deosn't mttaer inwaht oredr the ltteers in a wrod are, the olny iprmoatnt tihng is taht the frist and lsat ltteer be in the rghit pclae. The rset can be a taotl mses and you can sitll raed it wouthit a porbelm. Tihs is bcuseae the huamn mnid deos not raed ervey lteter by istlef, but the wrod as a wlohe. Amzanig huh? yaeh and I awlyas thought slpeling was ipmorantt!
It makes a big difference if your messed up words use common letter patterns (what, in the article he called 'Psuedowords'), or not.
Example:
'uesdnatnrd' wasn't to hard to recognize beacuase 'uesd' and 'tnrd' aren't letter patterns that exist in real words. So the mind works quicker to rearrange the letters to find a real word.
'aulaclty' was much harder because it's almost pronouncable. 'lac' and 'lty' are common patterns from real words, and 'aul' might not be common but it's pronouncable.
Just an observation.
Aw crap, ninjas!
I notice MS Research doing lots of basic research that has never been productized. Its rare to see corporations being so liberal with their resources. Even Google's very imaginative projects seem to be directed towards a commercial goal.
This suggests an interesting contradiction in MS product strategy. MS has a long history of "clone and conquer", e.g. Excel copies off VisiCalc and Lotus 123. Just this week MS cloned Apple iTunes. Yet MS Research is conducting some very interesting basic research. Go figure!
Following a day's sessions on legibility and word recognition, the Thessaloniki conference held a round table on legibility and the processes of word recognition, chaired by John Hudson and participants Mary Dyson, Hrant Papazian, Kevin Larson and Peter Enneson.
you had me at #!
Don't give any ideas to spammers on how to sneak their "pneis elnraegemnt ceram" past the filters. I do suspect that the effect is local to the small group of letters and long words that are totally randomized will be difficult to read.
The approprately clunky sounding phrase to express the thought related in your second sentence is "concatenative assemblage".
Furry cows moo and decompress.
One problem with deciding "word shape" vs. "letters" as the method a reader uses to recognize a word comes from treatment of the reader as "atomic". I am a proficient reader. When I read a word, whether written by another, or by myself (as I type), I have multiple subcurrents of consciousness. A typo in a word might leave me with recognition of the word, and a sense that "something's wrong", simultaneously - it sometimes takes me several seconds to detect the typo, especially if it's one I often make myself. Likewise, some spelling mistakes derive from the difference my spoken accent makes with the written conjugated spelling, most often in the case of syllables separated by an "e" that is pronounced as a "schwa", easily confused with some pronunciations of "i" or "a", and sometimes "y".
Reading words silently, I sometimes notice an inner chorus pronouncing the words, with one or two discordant notes, even from poorly organized structure or unparseable punctuation. Deciding how people recognize words must also account for how people's minds are organized. The myth of the "undivided self" gets in the way of understanding not only how complex we are "under the hood", where media is digested, but denies credit to our grand integrator, which juggles these partial selves into one face with which to confront the world. As machine intelligence benefits from multiple simultaneous processing, why should they have all the fun? As we mimic our own minds in computer simulations, why should we have all the fun?
--
make install -not war
I read really fast. I also read quite a bit of fantasy and science fiction. I have noticed the effect that weird alien and fantasy names (N'kalogh or Xyztle) are like driving over speedbumps. The higher the density of unfamiliar and nearly unpronounceable names, the more likely I am not to finish the book (or even pick it up).
"N'kalogh leapt onto his mighty huyloch and rode across the plains of V'looth'u". Next please.
This paper gives a convincing pyschological model about why this occurs and it is pretty much what I had surmised on my own.
So, from now on, please name all of your aliens Bob, Larry, Bubba, or Charles.
-- stream of did I lock the front door consciousness
ZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZ
Gentoo Linux - another day, another USE flag.
Here's one thing you might want to try if you do that. Get someone to track when you blink whilst reading and how often you blink, and whether it's consistent for a particular person, and compare it with other people.
:).
:) ).
It seems if you don't blink whilst reading it's like trying to eat food in big chunks... At least for some people. Then again it may be the stress of keeping your eyes open distracting you from reading?
My postulate is that the brain takes the blink time to dedicate more resources to processing and understanding what is read.
Coz it's amazing how much "brain CPU" vision/sight takes up for most people. I've got people (kids + adults) to try to do certain coordination tricks - like drawing a circle with the right hand whilst drawing a square with the right foot, or doing an OK sign with the left hand palm facing upwards, L sign with index and thumb of right hand, palm facing downwards, and then rolling hands over - switching to L with left palm down and and OK sign with right palm up, and back again.
I find that most people find it easier to learn how to do such stuff if they have their eyes closed and visualize it whilst doing it. Once they get it, they can do it with their eyes open.
Unfortunately many tasks that require high coordination require your eyes to be open
There seems to be a "brain state" difference between having my eyes open or closed. I find it harder to stand and balance on one foot with my eyes closed compared to with my eyes open EVEN THOUGH it is totally dark and my eyesight is useless, or I am blindfolded. (Of course it is much easier with my eyes open and the surroundings visible
I am not a scientist, and I haven't done a formal study on these items. So it's just anecdotal, but feel free to go do one - would be good if the resulting study is published somewhere on the internet.
That gives me an even better idea. Its probably much easier to detect when someone blinks then to track eye motion. So just change a few words of text whenever they blink. I suspect the result could be pretty funny.
As for why people blink at a specific rate, and whether that changes based on level of concentration, that's been studied.
"Studies have measured the blink rate and tearing on computer workers and noted that the blink rate dropped very significantly during work at a computer compared to before and after work. There was no significant change in tearing. The data support the fact that blink rate decreases during computer use, but also show that other tasks can decrease the blink rate.
Possible explanations for the decreased blink rate include concentration on the task or a relatively limited range of eye movements. Although both book reading and computer work result in significantly decreased blink rates, a difference between them is that computer work usually requires a higher gaze angle, resulting in an increased rate of tear evaporation. Since the main route of tear elimination is through evaporation and the amount of evaporation roughly relates to eye opening, the higher gaze angle when viewing a computer screen results in faster tear loss. It is also likely that the higher gaze angle results in a greater percentage of blinks that are incomplete. It has been suggested that incomplete blinks are not effective because the tear layer being replenished is 'defective' and not a full tear layer."
Which suggests that my blink rate may go up now that I have bifocals, since I have to look through the bottom of them for close-up work (lower gaze angle). I think a lot of the dry eye that I get is from the a/c, anyway.