September 16, 2003

visual word recognition

The interesting thing about blog entries is that you never know which one will actually interest people, taking off on a trajectory of its own, and which one will fade silently into the past. As for the whole jumbled letters in the middle of words not stopping the comprehension of a text meme, here's my followup entry with slightly more thought applied and (partially) in response to comments here and abroad in Blogovia.

First, here are the two texts. The first one appeared on September 12.

Aoccdrnig to a rscheearch at an Elingsh uinervtisy, it deosn't mttaer in waht oredr the ltteers in a wrod are, the olny iprmoetnt tihng is taht frist and lsat ltteer is at the rghit pclae. The rset can be a toatl mses and you can sitll raed it wouthit porbelm. Tihs is bcuseae we do not raed ervey lteter by it slef but the wrod as a wlohe.

The second text appear towards the end of April, and I blogged it on May 1.

Randomising letters in the middle of words [has] little or no effect on the ability of skilled readers to understand the text. This is easy to denmtrasote. In a pubiltacion of New Scnieitst you could ramdinose all the letetrs, keipeng the first two and last two the same, and reibadailty would hadrly be aftcfeed. My ansaylis did not come to much beucase the thoery at the time was for shape and senqeuce retigcionon. Saberi's work sugsegts we may have some pofrweul palrlael prsooscers at work. The resaon for this is suerly that idnetiyfing coentnt by paarllel prseocsing speeds up regnicoiton. We only need the first and last two letetrs to spot chganes in meniang.

David Harris was the starting point for the first text, and he also varied it a little to track how it spread through the net. He had received the original text through email on Friday, September 12.

Unscrambling the letters for the first text we get:

According to a researcher at an English university, it doesn't matter in what order the letters in a word are, the only important thing is that first and last letter is at the right place. The rest can be a total mess and you can still read it without problem. This is because we do not read every letter by it self [sic] but the word as a whole.

And, as I said earlier, and as many have pointed out, this seems like your typical bit of urban folklore. Especially vexing is the vagueness of "an English university" and no attribution for the "researchers". David Harris cleaned up the spelling a little, but left in the two word version of itself.

The reason, I tracked down Saberi and Perrott work as a possible origin is that Saberi is mentioned by name (and unscrambled at that) in the second, but older, text I had blogged back on May 1. Though, as many have pointed out, their paper (which I have yet to read as it isn't online) deals with spoken language and not the reading. Saberi took fragments from the middle of words and reversed them, but people were still able to comprehend them. If you search around with Google on such apt search strings as "cognitive science", "psycholinguistics", "visual word recognition", etc., you quickly find yourself reading a whole slough of psych papers. It seems that a lot of people have been working on all aspects of reading and listening comprehension. Times for word recognition are reduced when subjects have been primed with the semantic fields of the words. What happens to word recognition when the words are distributed in more or less random lists of a sentence? And letters aren't the only thing being scrambled: folks have been scrambling words in word order lax languages. Mark Seidenberg has an interesting paper on visual word recognition briming with saccadic eye movement, dyslexia, lesions, reading comprehension, orthographies, and Hebrew. And here's a paper on how the brain encodes the order of letters in a word that Wesley at Fear Everything pointed out in a followup entry to mine.

The bibliographies alone of these papers will provide hours of interesting lolling around on the web and reading.

[Addendum 09/16/03: I've read the Saberi and Perrott article. It seems that the chunks of speech they reversed came from everywhere in the speech chain and not just the middle of words. It seems then to have more to do with comprehension of the message in a noisy environment.]

[Addendum 09/17/03: See today's entry for a source of the first jumbled text.]

[Addendum 09/19/03: researcher and important are misspelled in the more recent text. I silently (and unconsicously) corrected them when I produced my clear text.]

Posted by jim at September 16, 2003 08:53 AM | TrackBack
Comments

For some reason, I am deeply trouble and perplexed that no one notices that "iprmoetnt" is spelled incorrectly.

Posted by: Jon Bruce on September 19, 2003 09:49 AM

Yes, researchers and important are misspelled. Were they in the original text? or did they get misspelled in the jumbling. You yourself left off the -d in troubled. I don't think it's much to get upset by.

Posted by: jim on September 19, 2003 10:14 AM

You might find the reaction from the artificial intelligence research people interesting. The implications to natural language recognition that can handle all of the typos and misspellings humans are prone to is significant.

Here is a page about a computer program that uses a similar approach to the one described in the original scrambled text. There is even a link to a free demo program does this.

http://www.ai-forum.org/topic.asp?forum_id=1&topic;_id=9528

Posted by: Ribald on September 19, 2003 10:36 PM

As an English teacher in France, and working with a lot of dyslexic children and adults, I very much appreciated your scrambling. So did my students!

Posted by: Tina Vartanian on September 23, 2003 11:41 AM

Here's a new twist to the "research" that's doing the rounds...!

Iltnsegnetiry I'm sdutynig tihs crsrootaivnel pnoheenmon at the Dptmnearet of Liuniigctss at Absytrytewh Uivsreitny and my exartrnairdoy doisiervecs waleoetderhlhy cndairotct the picsbeliud fdnngiis rrgdinaeg the rtlvaeie dfuictlify of ialtnstny ttalrisanng sentences. My rsceeerhars deplveeod a cnionevent ctnoiaptorn at hnasoa/tw.nartswdbvweos/utrtek:p./il taht dosnatterems that the hhpsteyios uuiqelny wrtaarns criieltidby if the aoussmpitn that the prreoecandpne of your wrods is not eendetxd is uueniqtolnabse. Aoilegpos for aidnoptg a cdocianorttry vwpiienot but, ttoheliacrley spkeaing, lgitehnneng the words can mnartafucue an iocnuurgons samenttet that is vlrtiauly isbpilechmoenrne.

Or, if you prefer:

Interestingly I'm studying this controversial phenomenon at the Department of Linguistics at Aberystwyth University and my extraordinary discoveries wholeheartedly contradict the publicised findings regarding the relative difficulty of instantly translating sentences. My researchers developed a convenient contraption at http://www.aardvarkbusiness.net/tool that demonstrates that the hypothesis uniquely warrants credibility if the assumption that the preponderance of your words is not extended is unquestionable. Apologies for adopting a contradictory viewpoint but, theoretically speaking, lengthening the words can manufacture an incongruous statement that is virtually incomprehensible.

LOL... it gets worse!

Posted by: Jeremy on September 25, 2003 04:40 AM

How thin the line between spam and urban folklore. Needless to say there is no department of linguistics at the Welsh university in question. And the page at the end of the URL is another scrambler proglet. Nothing in Google yet.

Posted by: jim on September 25, 2003 08:50 AM

I think a lot of the reason for this mental unscrambling is that the brain holds a kind of store of words. Most people read ahead a bit and put words into context. Speed readers do this even moreso. I would be interested to hear how dyslexic people get on with this unscrambling....

Posted by: Dave Lambert on September 30, 2003 02:23 PM

The order of letters in a word is important. What this example does not indicate is the context in which the words appear. Those of us who read this correctly were able to do so because we understood the context in which the words appeared from a grammatical standpoint. I bet you didn't read it very quickly the first time, but because we were curious we went back and read it again, but probably faster. It may be true that the eye does not attend to every letter, but it's also true that the eye does not attend to every word. In other words, we guess, a lot, and sometimes we guess wrong because of context. For example, I bet most of you would also be able to read this sentence and understand it if you do some guessing:

Babe _______ hit the ______ over the __________.

We can read this because we know the context in which it is written and we have the prior knowledge to fill in the missing words. But, what happens when we don't have that prior knowledge or we don't understand the context or recognize the grammatical structure. Obviously, we then attend to the word and all of its letters. That's what phonics teaches our students to do. How many times have you heard a student read something and use a word that was similar but incorrect? If so, he's probably not grasping the context or grammatical structure. Have you ever been reading something and gotten to the bottom of the page and despite the sentence being continued on the next, you read the first word on the next page before you've actually turned the page? I've seen students do it.

By the way, it's "Babe Ruth hit the ball over the fence."

Posted by: Mark on October 1, 2003 10:52 AM

I'm English, and I didn't understand the Babe Ruth example posted by Mark, because it refers to a game which isn't that popular in the UK, so there's a case that supports the idea that context is v. important.

Posted by: Looby on October 1, 2003 12:38 PM


On
http://lsto.gmxhome.de/hardlycrypt.html

you will find a three-lines perl script and a java script implementation to convert your text online.

Try it online!

Posted by: Daniel on October 14, 2003 02:06 PM

My only question is: Were the subjects of this study "taught" to "read" by look-and-say methods? Having listened to such victims of modern pedagogy read aloud, I have concluded that they don't look at the internal letters of a word anyway, only the first and last and then guess. Folks who were actually taught to read (that is, phonetically) do indeed slow down while deciphering this mess.

Posted by: Pogo on October 15, 2003 09:06 AM

Hey,
I'm an eighth grader, and my honors class is just starting the science fair. Several of my friends recieved an email from you on this theory, and I thought it would be an interesting topic for the science fair. You wouldn't believe how long it took me to track down this information! lol Now that I have found some details on this, I would greatly appreciate it if you could email me any more information. Thanks so much!!

Posted by: Alexis on November 21, 2003 07:41 PM

I really don't have any more information than is psoted here. Good luck on your project.

Posted by: jim on November 22, 2003 07:41 AM

Can someone please help me decypher this:

fof upi lmpe yjr nio;fomh mrcy fppt jsd nrrm fr,p;odjrf? gpt dp,r trsdpm yjsy tr,omfd ,r pg upi. O epmfrt eju?

Posted by: august on December 2, 2003 07:48 PM

I found this site through google. It has a very good explanation by a man who works in Cambridge University's Cognition and Brain Sciences Unit.
http://www.mrc-cbu.cam.ac.uk/personal/matt.davis/Cmabrigde/

Posted by: Claire on February 17, 2004 05:15 PM

august,
fof upi lmpe yjr nio;fomh mrcy fppt jsd nrrm fr,p;odjrf? gpt dp,r trsdpm yjsy tr,omfd ,r pg upi. O epmfrt eju?
translates to:
did you know the building next door has been demolished? for some reason that reminds me of you, I wonder why?

Posted by: Lincoln on March 15, 2004 01:15 AM

Me and my friend are doing a project for school, based on this topic, of how readers could still figure out words by not reading every letter. If there was any information you have available, or a number we could reach you at and interview you, please let me know by email.

Posted by: Sandra Trinh on April 1, 2004 10:18 AM

I blogged about this some time ago (here). One of the conclusions that I reached is that the phenomena of not reading every single letter/word has to do with the ability of some people to become adept "speed readers" and the theory behind speed reading in general.

I don't always "speed read" when I'm reading, but when I do, I can easily read a book or two a day. As I said in my blog post, perhaps when one is trained with speed reading techniques, one is better able to comprehend/decipher this sort of phenomena.

Posted by: Peter C. Frank on July 11, 2006 7:46 AM
Post a comment