How do we perceive and read words and sentences? Does the order of the inner letters play no significant role?

Question

Try to read this texts, start with the most difficult one, if you cant read, skip to the next easier one:

all letters mixed

I onlucd't ieebvel ttah I udloc talyulac rsddetanun hwat I swa radgeni. sngiU eth inrdebelic rweop fo hte nmauh imdn, dorcincga ot arceehsr ta meaiCrgbd yvisUntier, it deson't ramtet in athw rored eth erlttes ni a rwod rae, hte noyl atomirnpt nhgti si ttha hte tsfir and last letert eb in het girth leacp. eTh sret anc be tolta ssem dna ouy cna litsl eard it wohutti a elrpomb. ishT is secbaeu het mahun mnid dose tno aerd verye leettr yb efsilt, btu hte orwd sa a leohw. agzmnAi, huh? Yhea nda I awlyas utgtohh lgniespl aws optnrimta! eSe if uoyr resfdin acn eadr siht oto.

inner + first letter mixed

I dloucn't ebeilve taht I lcuod utllaacy tsnredanud awht I was irdenag. sUnig hte eidbcrline poewr of the umhan nimd, ccrdiaong to aecrersh at biCagmdre eivUnirsty, it esdon't eattmr in hawt droer the teltres in a rwod are, the lnoy ropamintt nthig is that hte rfsit nad alst eltetr be in the hgrit aclpe. The rest acn be ttoal mses and oyu acn tilsl aerd it ouithwt a rebplom. hiTs is ecuasbe hte muahn mind does ont ared every leettr by iletsf, but hte orwd as a hlwoe. maiAzng, uhh? Yeah nad I lwyaas hugthot lisneplg was mprnotait! eSe if oyur nridfes acn raed iths oto.

inner letters mixed

I cnduo't bvleiee taht I culod aulaclty uesdtannrd waht I was rdnaieg. Unisg the icndeblire pweor of the hmuan mnid, aocdcrnig to rseecrah at Cmabrigde Uinervtisy, it dseno't mttaer in waht oderr the lterets in a wrod are, the olny irpoamtnt tihng is taht the frsit and lsat ltteer be in the rhgit pclae. The rset can be a taotl mses and you can sitll raed it whoutit a pboerlm. Tihs is bucseae the huamn mnid deos not raed ervey ltteer by istlef, but the wrod as a wlohe. Aaznmig, huh? Yaeh and I awlyas tghhuot slelinpg was ipmorantt! See if yuor fdreins can raed tihs too.

Actually we learn reading by putting together letters, but with ongoing time this seems to change drastically. Do we perceive words and sentences as small pictures? Like mosaics build of morphemes? Is psycholinguistic research done on this topic?

---

on comments: im very interested in Psycholinguistics or psychology of language. Wether this belongs to linguistics.SE, i dont know, the FAQ is unclear, I just read this on-topic question about perception and was reminded of this phenomenon, thougt some experts in linguistics can shed some light on this imho important & interesting phenomenon and its implications as a method, topic for linguistic research.

Maybe it is mainly explored in psychology faculties. But as a scientist myself, i dont see how you want explore origin and evolution of languages without analysing how we perceive (see/hear) language, how our ears and eyes are limited in perception,memorizing and remembering of frequencies and micro-shapes. Thats like saying, we can explore music by analysing all note-sheets ever written or understand PSE by studying chemical compounds instead of doing atom physics. This would be unscientific and missing the big picture, maybe missing the main factors. Of course you have to start with psychoacoustics, phonetics. Maybe you learn more about origin of human language by studying differences to other animals (singing whales, birds) than analysing the fine nuances of human spoken languages.

So im a bit surprised that so many see this question as off-topic, despite how interesting (to me) the current answers and their links & implications to different linguistic fields are. If a whole PHD-thesis covers this topic, how can it be off-topic? But i wasnt commited to this site and its definition, i made a question in meta, imho you should move discussion and comments there before i or somebody else wastes time writing a perception-tagged closed/non-voted questions yielding less effort.

Answer 1

Matt Davis provides a useful summary related to this meme. Versions in Spanish, French, Dutch, Danish, German, Czech, Icelandic, Portuguese, Swedish, Bahasa Indonesia, Russian, Albanian, Hungarian, Italian, Irish, and Polish are available on the page.

Hebrew and Finnish are also linked and Davis mentions:

[of Finnish] Both Peter and Ari suggest that the resulting scrambled text is very difficult to read. ...one thing that makes these scramblings difficult to read is that the jumbled-letters often move across morpheme-boundaries.

[of Hebrew] Samuel suggests that the scrambled text is "a REAL mess.. you can not understand it at all". This may reflect a interesting property of the Hebrew writing system. Since vowels are not written in the text, there is a lot less redundancy in written Hebrew. It may be that readers are already using some inference processes to work out what words are written, the extra load added by jumbling letters creates an additional, excessive level of difficulty.

Davis mentions that while no research that spawned this meme was done at "Cambridge University" (some versions of the common text say English University, in fact), this text may have been more widely circulated after a 1999 note from Graham Rawlinson containing jumbled text, related to his 1976 PhD research that described how readers can understand text despite high noise or error rates.

Languagehat also discussed this in 2003, and many of the commenters there have good points about what to explore further. Here's one, S.W.:

Most of this seems fairly close to the concept of non-words and orthographic non-word neighborhood effects (i.e., RT's in a yes/no word/non-word lexical task being longer when non-words have a higher neighborhood, or lexical inhibition). Since most of these are strictly non-words, I'm betting that the non words with higher orthographic neighborhoods are harder to read.

Also, since we're talking about representing these non-words as actual words I'd be willing to bet there are two very obvious factors involved. I'm betting the base-words (non scrambled) with a higher orthographic neighborhood are even harder to recognize, e.g. high non-word orthographic neighborhood, high base-word orthographic neighborhood = much more difficult to recognize.

And, as someone pointed out, we're not even considering ascenders, descenders, syntax, schemas for written text, the degree to which each word is scrambled....

I'm not aware of what research covers those issues but it mgiht mkae for a good porjcet or fololw-up qeutsoin.

Answer 2

I couldn't read the first and not the second much either.

So, why the third one is simpler? There was a study done by the Cambridge University, that stated that the order of the letters was not important, except for the first and the last one. If you mix them all, like in the first two, you'll need time to anagram them and proceed, but with the last one you can read it faster (and more easily).

The neologism created to denote this cognitive process is Typoglycemia, a portmanteau of typo + glycemia.

The origin of this seems to come from a letter sent to the New Scientist (I can't find a version that can be read without subscribing), by Dr. Graham Rawlinson, although it also seems that a researcher think it was wrongly-attributed (read here). I think you can also check his "The significance of letter position in word recognition" in 1976.

Not all languages seem to work like this. Arabic for example, because in the written form it has no vowels (you add them reading, if I remember well); Finnish and Turkish words are too long, and for languages like Thai that has no space between words, it doesn't work.

Lastly, here is a site where you can type or paste text and mix it (the button Mischia! does it).