Letter Swapping Meme Sparks Discussion of Word Shapes

I think this research is a little flawed though, as it still requires knowledge of the original word, something that children would not have. A non-native English speaker would be rather screwed, same for say visually impaired readers.

Anyone tried this in another language - French, German? or put this through a screen reader?

I would like to think Russell Holban got there first with his book, Pilgerman as a possible scenario for something like this, even if it was more phonetic.

That said it does crack open a debate around language development in a population/ demographic, the acquisition and perception of 'word' symbols, all of which have long been argued.

I see that the discussions elsewhere are pointing to a possible hoax, which is yunnf, we all abreviate and adapt language to situations especially in say online 'chat mode', seems like an idea to pursue further.

Great post, Colin. I myself haven't been able to find much on the topic so thanks for the links!

You know what makes me really happy?
That "bouma" finally seems to have
become a Real Word.

> It would be great if Graham Rawlinson’s
> dissertation were made available

I'm tracking it down as we speak.

--

It's not the sum of distances that a letter has moved or anything like that - simply because we don't read by compiling letters. It's a matter of external envelope: the leftmost and rightmost profiles, along with the stuff sticking up and down from the middle (ascenders and descenders) are more important than the x-height stuff in the middle, because we recognize shapes outside-in. This is pretty well accepted. What is under contention though is why/how this happens. I think it happens because we process successively higher-resolution (but slower-to-process) layers of what's in vision, and we stop at the layer that gives us enough (not total) certainty of what the bouma is. The threshold at which a given bouma is adequately recognizable depends on many things, especially semantic context. But first and last letters are indeed generally more important than the middle letters, as both empirical and anecdotal research has shown.

hhp

Just to answer the question about other languages. I found this somewhere in the net and as a german speaking person I have to say that it is all clearly readable...

"Gmäeß eneir Sutide eneir elgnihcesn Uvinisterät, ist es nchit witihcg in wlecehr Rneflogheie die Bstachuebn in eneim Wrot snid, das ezniige was wcthiig ist, ist daß der estre und der leztte Bstabchue an der ritihcegn Pstoiion snid. Der Rset knan ein ttoaelr Bsinöldn sien, tedztorm knan man ihn onhe Pemoblre lseen."

I've read this text in English, Spanish and French. And I've been told by a Danish speaker that it also exists in Danish.

So I guess it's possible to write a text easily understandable even if we move some letters.

You konw, prat of it is taht it's not so easy to mcuk wtih srhot wodrs.

Ayanwy, for thsoe so inlcnied, hee'rs a litlte Pyhotn spcirt whcih you can use to test the same efefct.

As this is about recognizing word shapes, I think Marc is perfectly right in saying that you must have a knowledge of the word beforehand (in its written shape). Given that, it doesn't matter if you are native or non-native. Being German myself, I can read scrambled texts with perfect ease in German, English, and even Dutch, a language I only read, but cannot speak at all.

Ayanwy, for thsoe so inlcnied, hee'rs a litlte Pyhotn spcirt whcih you can use to test the same efefct:

Hey, cool! It works.

This whole idea will be disputed at ATypI whe n Kevin Larson of Microsoft speaks on “The psychology of word recognition”.

Hey will there be a online copy of the Kevin´s speech ?

So, thanks to Jim Bisso, I have been found and obviously asked to reply to the above.

First, I am not sure when Russell Holborn wrote Pilgerman (sorry about that) but my research was 1973 to 1976, and was probably a sort of first in that to conduct that many experiments you needed a comptuer controlled display, and computers were only just able to do that kind of thing unless you had loads of money (I did it with priunted text and screen read subliminal recognition techniques).

As to word shape, sorry, but I went through that throroughly, and when you replace ups and downs with other ups and downs you lose easy recognition. We need the actual letters even if they are in the wrong up down order. Shape might count for something but not a lot.

E.g. prtamgaic or prdenyiic, or even (to give shape a chance) preynohic

So which one reads, pragmatic?

Yes, you have to have knowledge of the letters of the word, of course, but I found young children soon were able to do this kind of recognition, good readers from 7 or so.

I did do some rather primitive calculations about processing time, (don't forget, I was subiminally displaying words so I got a lot of (did you get it or did you not kind of information) and my conclusion was that there must be some kind of parallel processing of letters ireespective of position, which would kind fo make sense.

As to Mike Larson of Microsoft disputing this kind of thing (a big generalisation) I would be pleased to hear from him).

My thanks to Jim Bisso for tracking me down.

More on what I have done can be found on a Google search with my name.

Or Just Email me.

Graham Rawlinson

Or is it Ralnwsion?

> parallel processing of letters
> ireespective of position, which
> would kind fo make sense.

But it wouldn't explain the primacy of initial and terminal letters. The only thing I know of that does is the theory of word shapes (boumas). But I will note that you have to be reading very "normally" to benefit from the speed you get when you stop looking at all the letters (which you can't do in the parafovea anyway).

Dr Rawlinson, I was about to start a search for your fascinating paper through UCLA's Inter-Library Loan system, but since you seem to be so accessible would be too much to ask for a copy of the paper directly? That would be wonderful.

hhp

> So which one reads, pragmatic?

I see three problems with this test:
1) It is devoid of context. Word shapes are never enough in a vacuum, they're used to speed up the choosing of a possibility out of contextually valid ones - they can't make you know a word out of context when most of the letters are changed, even if extenderness is maintained (which the "y" doesn't even do very well, btw). Reading is all about efficiency, not absolute "correct" answers.
2) During immersive reading you don't have many seconds to figure out what a word is (or was, if you're only shown the word for a very short duration but then given time to figure it out), you have a fraction of a second to be sufficiently confident, so you can move on to the next set of boumas. One thing this means is that you don't have time to shuffle/swap letters in your head and eventually find a match deep in your lexicon.
3) During immersive reading you use the fovea less than the parafovea, this latter having not enough resolution to convey the innards.

So this type of test is great for finding out what happens in deliberative foveal processing, but not immersive reading. In a way, it's an algorithmic task, versus the heuristics of normal adult reading.

In contrast, the test paragraph that Colin posted is where first and last letters do more of the "work" than the middle ones (as do extenders).

BTW, there are two other factors one must not forget:
1) A major determinant of a bouma is length. So for example subbing an "l" for an "h" throws things off.
2) The issue of "half-ascenders", like "t" and "i".

hhp

BTW, some of you might have already seen this, but:
http://www.the-djs.com/~joenix/uncorrect

hhp

Regarding Kevin Larson's upcoming ATypI presentation, I don't know exactly what he intends to present, but based on previous discussion with him and knowledge of the kinds of studies he has been involved with, I suspect he will focus on how individual letters contribute to our perception of words. This doesn't seem to me to be disputing what I understand of Dr Rawlinson's findings. Indeed, if, as stated above, 'when you replace ups and downs with other ups and downs you lose easy recognition', that suggests that although the initial and final letter are extremely important the actual presence of the correct intervening letters, even if out of order, is important to recognition.

This is a fascinating discussion. As an applied cognitive psychologist, I'll be interested to see whether Rawlings' dissertation turns up. Either way, these findings really don't add anything to what we already knew in the 1960s about the effect of contextual cues on word recognition. As Colin Hartnett observed above, readers' ability to recognize words is fundamentally influenced by their recognition of the context of the words in sequence. If we reconstructed sentences in random, jumbled up order would readers still recognize words? Absolutely not. Anyone interesed in the seminal research on the effect that Rawlings has possibly "rediscovered" should take a look at Tulving, E. and Gold, C. (1963). Stimulus information and contextual information as determinants of tachistoscopic recognition of words. J. of Experimental Psychology 66: 319-327. I do hope that Rawlings explained his findings in light of Tulving's and Gold's work. I would love to see the dissertation and will re-check this website to see if anyone has uncovered it.

Those of you interested in how non-native English readers would perform given a paragraph like the one above need only try it on yourself with nonsense words. If you don't read the language well, you don't get as much contextual information from the word strings and will fail to recognize words that a native reader would recognize. Here interested readers might just look at even earlier research by Miller, Bruner and Postman (1954) that demonstrated that more familiar we are with words, the less time it takes us to recognize them. Their work is published in:

Miller, G., Bruner, S. and Postman, L. (1954). Familiarity of letter sequences and tachistoscopic identification. J. of General Psychology 50: 129-139.

I love typography, but I don't think the shapes of letters is going to help us explain the meaning that words convey.

Good discussion!

Sam

> although the initial and final letter are extremely important

But why would they be so?...

> the actual presence of the correct intervening letters,
> even if out of order, is important to recognition.

Although I can't be confident until I read Dr Rawlinson's research, I have a hunch as to what's going on: since word recognition proceeds in stages (starts blurry, get sharper as necessary), if you give enough time to the reader to figure out what the word is/was, wrong letters (as opposed to displaced ones) will pull the handbrake on recognition down the line, but there's no time for this to happen during immersive reading. But boumas still help anywhere.

So yes, none of the letters in a word are unimportant, but the more they're in the middle (and generally the more they're x-height) the less important they are in immersive reading, as opposed to tests that are actually deliberative.

--

> I don't think the shapes of letters is going to help
> us explain the meaning that words convey.

This is interesting - could you elaborate?

hhp

So we have (what we could henceforth call) the "msesgae":

"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".

Even if it’s funny, this "msesgae" is an improper and excessive generalization, which conveys an extremely reductive vision. Moreover, whereas it should only remain what it is, i.e. a simple fantasist and entertaining text, it is taking worrying forms (we see it in mails, weblogs, chat-rooms where participants, absolutely amazed and amused, are venerating this "sensational discovery" and friends from everywhere (also excited) are forwarding it in different languages (apparently, this “hoaxmeme” (hoax + meme) is floating all over the web).

Let’s try to encircle the topic (not by haughty pedantry but just by anticonformism and anti-“simplistism”). If you were looking for a serious explanation of it, here is an “anti-hoaxmeme”:

Introduction
Reading is a complex activity that involves many aspects of knowledge, which are of various natures and various complexities (this is due besides to the fact that “writing” is complex). It's an activity, which implies cognitive processes but also, simultaneously, perceptive processes: reading, it's to perceive and to identify words.

Development
Many linguists worked on the description of the mechanisms’ evolution of the words’ identification and there are now many developmental models of reading. The principal models comprise three way of reading, which correspond actually to three chronological stages of acquisition (for this presentation, let's start with the second one):

- the alphabetical reading (second stage): the reader connects the oral examination with the writing (in other words, he learns how to make correspondence between letters and sounds (ex: the sound [k]can be written with 'c' (cot), 'k' (kiss) or 'ch' (chord)). At this stage of phonological mediation, there is a code training; the learner enriches its phonological knowledge and transfers it to new words (it’s a form of self-training). This stage is called an "indirect way" because the reader reads the words through a decoding process.

- the orthographical reading (third stage): the words are analyzed in orthographical units (orthography indicates here the sequence of letters forming the word). There is no phonological conversion; the words are read and recognized directly in reference to a memorized orthographical lexicon. This stage replaces gradually (but not entirely) the alphabetical one. The reader does not need to decipher anymore: he recognizes the words through a "direct way".

- the logographic reading (which is actually the FIRST stage in the reading training): at this stage, the reader uses various kinds of clues to 'read' the words, inter alia, those provided by the extralinguistic environment. The letters’ order and the phonological factors are not taken in account, but the visual clues are. There can be at this stage an instantaneous recognition of familiar words (or somehow ‘learned by heart’), and the riddles made on the basis of projecting visual clues allow the constitution of a first total vocabulary. The visual clues can simply be the length of the word or its "silhouette" (outline) or even just one letter. The classic example to illustrate this stage is the word: "Coca-Cola”, of which logo is easily identified by almost all children of 5-6 years old. If we change only one letter of the word: “Coca-Coca”, children will not notice the difference from the original word (adults neither sometimes, as some experiments proved it).

The most perspicacious of you may have already understood: what occurs actually when we read the "msesgae", it is that we, literate readers to whom reading and writing have been taught, use our competences, acquired and automated thanks to years of reading experience. In other words, we have developed "HABITS" of reading.

The "msesgae" experiment could let us think that we get back to a logographic reading, in which access to significance is carried out directly via the pictorial semantic system (with words treated like images-logos), but this is not completely true.

Actually, we continue to use the orthographical reading system (in which access to significance is carried out via the verbal semantic system). If we look at the "msesgae » more closely, we can notice that 34 of its 68 words (short and common by the way), are correctly spelled (50%, half of the text, and most of them are "grammatical words"). Added to a simple and common syntax (journalistic style of the “forma brevis”) and our capacity of anticipation and auto-reflex correction of more or less experienced reader (the system used is close to the "typing error" one, and anyway, teachers manage quite well to read our essays stuffed with spelling mistakes. In other words, you don’t have to be a Professor of literature to spot "what" in " waht "!!!), it gives many visual clues!!! (Moreover, there is a syllabic facilitation phenomenon, but I skip the details).

Conclusion
The proposition, which is conveyed through the «msesgae», is not completely false but it is very reductive, and completely incorrect when it affirms that only the place of the first and the last letter of the words do matter. Actually, it deals more with their "silhouette" (from which our (almost standard) system of abbreviations rises (another facilitating clue)). If we can read the "msesgae" without any problem, it is because we are good readers reading a text easily accessible in spite of its orthographic and spelling mistakes.
To prove it, if I give you the correctly spelled words "acetoxybutynylbithiophene deacetylase" or "carboxymethylenebutenolidase", dear expert readers, you will resort to an alphabetical analysis (second stage) and will use a grapho-phonological decoding for these unknown words (I suppose, this experiment may not always work if you are chemist, druggist or doctor... if it’s the case, sorry for this affront :-).
Another counterexample: if you read AT THE FIRST GO the following sentence as quickly and fluently as you did with the "msesgae", all my theoric explanation goes down the drain (or you are an innate champion of anagrams!):

“Nreuuoms pmeeononnhs peossss uiapocmltecnd etaaoilxnpn; nwttdtsniinoahg, the pdseuo-snfiiiectc spssliiimtm is not snfiiiectc and eieecndvs are oetfn mdanleiisg”*.

Guillaume Fon Sing,
(alias GUITCHUS)
guitchus@hotmail.com
Linguist

* “Numerous phenomenons possess uncomplicated explanation; notwithstanding, the pseudo-scientific simplistism is not scientific and evidences are often misleading”.

Please forward it, …it can teach sb a thing or two.

There seems to be some confusion about
this ’english’ university.

Is this it?

www.warwick.ac.uk/~pssbab/papers/lcp.pdf

ƒ

Guillaume, why do you think some people have trouble accepting the silhouette model? Besides thinking about reading itself, I tend to consider "meta-issues" like what people think about reading, and I've come to the conclusion that most people who reject it certainly don't do it because they have the knowledgable background, they do it because of some sort of internal belief system. If this is true, it might be because of a desire to give consciousness (deliberative reading) the upper hand over the subconscious (immersive reading), in a way to gain control over Nature, this sentiment being a classic western weakness: obsession with Control. I think many people who cling to the old and false letterwise model are basically afraid to admit the primacy of the subconscious, since it's something they can't control.

hhp

>TRADUIM:
>I've read this text in English, Spanish >and French. And I've been told by a >Danish speaker that it also exists in >Danish.

In danish the text is allmost unreadable if you only leave the first and last letter untouched. Two letters in the start and in the end needs to be the same to obtaine the same readability, as in english.

-Nicolaj

Snopes has listed another more recent article , called "An anatomical perspective on sublexical units: The influence of the split fovea," on their site.

That link leads to the exact same PDF as mine two posts up.

I just don't know how to do those fancy highlighted links.

ƒ

Fredrik, that paper is worth a read, just seeing how recent it is.

hhp

Fredrik: Sorry, but the page was cached and the earlier version didn't have your comment on it. As soon as I'd posted, I saw the uncached versionof the page, and I had that "d'oh!" feeling. Sorry about that.

Thanks for discussing.
Beeing completely new to the topic it just leaves me breathless.
Though it is not close to the main: Is there anything similar in music?
Let me read more.
Michael, Germany

I have recently posted the following on the Cogling discussion group which I thought would be relevant to this typographica discussion on the now famous ( or infamous) Graham Rawlinson passage described in the above article by Colin Hartnett:

"It seems the question arises what would be the result if the last and first letters of the words were likewise scrambled. The last and first letters serve as word boundary markers that are intricately related, both phonetically and semantically, to the internal letters of the word. However, if we were to scramble word boundary constraints as well as the internal letters, we would also have to scramble the white space appearing between words which , also, marks the boundaries between words. In so doing, I think it should be obvious that we would arrive at a final jarbling and mixing of letters that could not be easily read, but nonetheless, could be deciphered with some effort.
However, the point I wish to make is that a true scrambling can never be achieved, even using the most refined computer randomization techniques. Scrambling of letters does not scramble the mind and our innate phonemic notions of feet, phonological markings at word boundaries, coda and onset requirements, etc. Deciphering scrambled letters, then, is not simply an exercise in pattern recognition or Gestalt psychology. Rather, it is primarily an exercise of applying phonemic rules to letters that in their scrambled state violate those rules. Here, I think the conclusion of the researcher of the scrambling experiment is wrong. His experiment does not demonstrate we do not need to read individual letters within a word. This is aside from the issue whether we do or we don’t when the letters are in their correct sequence."

---Asa M. Stepak

> applying phonemic rules

Does this mean you think we use sounds during immersive reading?

hhp

If the letters were out of sequence or we came across a strange word we would have to apply phonemic rules. Otherwise, I believe we typically do but to a much lesser extent. The whole point is letters are symbolic of the underlying phonemes which represent conceptions of the mind. This is why I stated letters cannot be absolutely scrambled. The position of a letter in a word and the letters adjacent to it affect our mental conceptualization of the letter. This would be more obvious if we scramble the underlying phonetic segments represented by letters. On a phonemic level, it is well known that the particular sequence of phonetic segments determine the overall phonetic aspects of that sequence. Thus, if you switch one phonetic segment from one position to another you essentially change the phonetic segment you are switching to conform with phonemic rules. Thus your attempted scrambling or randomization is constrained by phonemic rules and you cannot achieve an absolute randomization.

> If the letters were out of sequence or we came across a
> strange word we would have to apply phonemic rules.

The latter, always yes. The former, usually no.

The reality of immersive reading is that we don't read (ie need) all the letters, and this precludes high phonetic relevance. This has been known for decades, and is in fact directly derivable from the research of Emile Javal, done over a century ago.

hhp

Well, I just don't think this is something that can be proved. Letters are not phonemes. So you can leave out a few letters and still know the underlying phonemic content such as in semitic languages that rely solely on consonants. Similarly, I would not dismiss phonetic relevance as being important in immersive reading. We cannot separate and detach our innate phonetic conceptions of words from our understanding of them. Phonetic content has a relationship to semantic content. There are instances when need not have to read all letters since there are other ample cues, such as context, that allows us to comprehend the written passage. But I don't think anyone can show or prove that we can comprehend a written passage without , also, comprehending the underlying phonemic content on a conscious or visceral level.
Letters are there to serve as cues but can become surplasage if there are ample other cues. No one ever said written language was a perfect invention. In this regard, it would be wrong to equate letters as having a one to one relationship with phonemes.

Does this mean you think we use sounds during immersive reading?

Among the many interesting studies referenced by Kevin Larson in his ATypI presentation on the psychology of word recognition, is one in which subjects were slower to correctly identify words which are not pronounced the same way as similarly spelled words. For example, for the group of similarly spelled words 'hint, lint, mint, pint', subjects were slower to correctly identify 'pint' than the other words. This is interesting, because it indicates that our brains are making connections between writing and the sounds of words, even if this information is unhelpful.

It is my understanding that immersive reading is not 'speed reading' where one actually decides to skip (ignore) certain sentences or groups of sentences that appear to be redundant or not germaine to the overall content of the passage. Clearly, if one chooses to skip a sentence, or two or more consecutive sentences, the underlying phonetic content of those sentences is also skipped. However, the boundary line is not clear and distinct. You can skip a few letters or skip a couple of words or part of a idiomatic phrase and still have a pretty good sense of the underlying phonetic content. This is partly due to the fact that words and especially frequently used words have collocation relationships with other words. Letters also have collocation relationships with each other determined by phonemic rules and semantic constraints. Thus, there is no rigid dividing line determining how much visual contact with letters is needed to maintain a sense of the underlying phonetic content. Complete visual contact is not needed but some degree of contact would be needed which would probably vary from sentence to sentence.

Asa, from what I understand the average length of a saccade is 7-9 letters, and this doesn't vary very greatly. In other words, immersive reading tends to be pretty regular, although saccades appear to be at the longer extreme if short words are involved. Another of the studies referenced by Kevin Larson indicated that readers scan about 15 letters ahead of a fixation*, i.e. about double the length of the average saccade, and us this information to program the location of the next fixation. In other words, the length of the saccade is determined by the information retrieved from parafoveal vision.

*The study measured reading speed changes when letters ahead of the fixation were blocked out. Reading slowed when letters less than 15 characters ahead of the fixation were blocked out.

By the way, due to massive demand from attendees of the ATypI conference, it looks as if Kevin is going to write up a summary of the studies referenced in his presentation. A brief report of the presentation and an extensive reading list are here.

Will the Gerald Unger talk on legibility, or a summary, also be available?

Saccades average 10 characters, but in fact vary greatly. Sometimes they go to 15. Coupled with the fact that we can't differentiate letters outside the fovea (3-4 characters), the conclusion is pretty simple.

Of course, the worse the reader (or reading conditions, such as ClearType...), the shorter the saccades, and the more one relies on individual letters.

hhp

As the infamous Graham Rawlinson whose research has been widely (and wildly) quoted, I would just like to add a couple of thoughts and corrections.

First, I did not suggest that we do not take in individual letters, rather the reverse, my hypothesis after the research (so it would have needed more research) was that the individual letters were 'recognised' in a parallel processing way, to you have a system which can fast match for letter content. What I suggested was challenged by my results was the predominance of word shape theory.

Second, I also found that having the fiorst and last two letters made the reading almost unnoticeable, for fast and not prewarned readers anyway. If you did only first and last letters it was fairly noticeable and for some quite distracting.

Interestingly I started my research trying to look at eye movements, like Larson of Microsoft. I would love to correspond with him as he seems to have followed a similar path.

Does anyone have his Email?

I did do some 36 experiments trying all kinds of combinations, vowel jumbling, replacement with x's and all kinds of things. As a PhD student I was mroe interested in raising questions by innovative research methods (computers were fairly new at the time) rather than suggesting I knew the answers.

Thanks for listening

Graham

It turns out that UCLA can't get a copy of Graham's paper for me through the normal channels; I'd have to submit a purchase request for it, and that can take 3-6 months... Graham himself only has one personal copy apparently, and I wouldn't expect him to photocopy the whole thing for me and mail it out... So if there's anybody who knows anybody who has access to that Nottingham U library and is willing to copy it for me, maybe we can do some kind of barter?

hrant{thesign}inverselogic.com

hhp

Obviously our ability to read the meme is made faster because of our ability to decipher the meaning of the word in it's given context. If words are taken out of context the brain takes a much longer time to process the scrambled letters. example: the scrambled word "dmaned" . I have kept both the last and first letters in place and scrambled the middles. This word could be interpreted as "demand" or "damned". The word shape also does little good in deciphering the scrambled word. Does this apply to all words?

> The word shape also does little good in deciphering the scrambled word.

The bouma (word shape) can be seen as consisting of: length, lateral profiles, extenders, and internal details. When you're foveating (ie deliberating) on a single word, nothing does much good. But in immersive reading, the coupling of context and bouma information is very valuable, especially to experienced readers who have a large "lexicon" of boumas and the ability to decipher them deep into the parafovea. If your example ("dmaned") is encountered outside the fovea (which is statistically likely, at least for a proficient reader), then: the bouma length will limit the possible choices of word nicely; the lateral profiles aren't too great so they won't help much; the two ascenders are helpful especially since they're close to the ends; while the internal details aren't too hot. If all of this information coupled to the context are not enough to adequately map a word to the bouma, then you fixate on the word. So obviously, the following helps readability: divergent letter widths; divergent lateral profiles; divergent extenders; and last and least, divergence in the x-height. However divergence can be taken too far if a letter is made to look "alien", causing an errant fixation to it from an earlier (and sometimes later) line.

hhp

The use of boumas is not fundamental to reading. And this is particularly true for "remembered" boumas.

Consider a unicase alphabet, such as Bradbury Thompson's, where he replaced A,E,M,N and R in an all-cap typeface, with their lower-case forms.

What's left has very little boumacity, other than word-length, and yet it can be read quite smoothly. It's possible to create other, more unusual unicase fonts, using standard letterforms, but with different combinations of upper and lower case glyphs, and they still read surprisingly well, despite the fact that the reader is dealing with virgin boumas.

Marvin Minsky's theory of music -- that listening is a heuristic experience (ie the listener's task is to start listening, and "learn" the structure and meaning of the piece as it progresses) can be applied to reading. It is the reader's task to figure out the set of rules that will enable the writing to be decoded and understood.

Certainly, familiar codes will be deciphered more quickly, but that tells us very little about the decoding mechanism, which is many degrees smarter, and flexible enough to handle every written and spoken language in the world, as well as Braille and sign language.

About all one can say is (as you say, Hrant) that the glyphs should be divergent in form, to aid differentiation, but not so divergent as to impede flow. Deciding what this multi-dimensional balance point should be is a job for artists and designers, not scientists.

I'm not suggesting limits to the scope of scientific research, but I do object to the "paint by numbers" engineering it can lead to.

> The use of boumas is not fundamental to reading.

You're talking legibility, not readability. For the former even FF Extra is fine. But don't try setting a book in it!

So it's a matter of efficiency, not absolutes.

hhp

>You're talking legibility,

Of course. This IS the "letter swapping meme" thread.

Sounds like you're admitting that boumas have little to do with legibility and word-recognition, but improve readability...

But this "letter swapping phenomenon" is all about boumas as far as I'm concerned. Specifically, the point was to explain why the first and last letters are more important than the middles.

Sure, when you have all the time in the world boumas are moot, because compiling totally decipherable letters leaves no room for doubt. But immersion is all about doubt.

hhp

>the point was to explain why the first and last letters are more important than the middles

That's easy. It has nothing to do with the visual shape of the word.

Here are two contributing explanations.

1) Readers are used to hearing, mentally cataloging ,and identifying words starting with the first letter. An anagram of the 7-letter word "According" is much easier to identify knowing that the first letter is A, rather than the 2nd letter is c.

2) If a letter is fixed in position at the beginning or end of the word, there are fewer possible syllabic combinations with the other letters in the anagram than if it is in the middle, with letters on either side.

Nick, this phonetic stuff holds no water.

hhp

It's not phonetic.
It's about confirming the near-as-damnit spelling of a word based on recognizing the letters that make it up. Syllables, such as "ing", are useful building blocks.

Yes, "ing" is certainly a great bouma. :-)

hhp

"ing" is a sequence of 3 individual letters that appears frequently in English. If you are trying to solve an anagram (or make a word in Scrabble) and you have n, g, and i in your puzzle, with the last letter g, then "-ing" is a good bet.

This is the way the letter-swapping trick works -- it can be explained as a spelling puzzle pure and simple -- without resort to reading theory.

Not when you consider that 2/3-rds of reading is done in the parafovea where things are too blurry.

What is your own theory of reading?

hhp

The point I'm making is that the letter-swapping trick works because words are spelled with a sequence of letters.

Whether the letters in the anagrams are perceived individually or in groups, blurry or in focus, they have to be sorted enough by the reader to identify the correct words they represent. It doesn't matter how the reader does the recognition and sorting, the configuration of fixing the first and last letters makes a puzzle that is easy to solve for the fluent reader.

I don't subscribe to any current theories of reading. When they start talking quantum holograms, then I'll climb aboard.

"A text is a succession of words, but
a word is not a succession of letters."
-Kolers

Immersive reading (as opposed to looking at posters) is heuristic, not algorithmic - just like anything else in the subconscious. You don't need all the letters.

> I don't subscribe to any current theories of reading.

?
It's funny, I thought you were trying to explain the phenomenon...

Are you in fact a closet "We Read Best What We Read Most" guy? It's nice that they used to prance about, but now they're much quieter.

hhp

>You don't need all the letters

That's why I said "enough" for recognition, in my previous post.

>I thought you were trying to explain the
phenomenon...

I was answering your question "why the first and last letters are more important than the middles," by saying that it was a matter of solving a spelling puzzle -- WITHOUT recourse to reading theory.

>Are you in fact a closet "We Read Best What We Read Most" guy?

Like I said, I don't subscribe to any theories of reading, closet or otherwise. I would certainly agree with (Ms Licko's?) quote most of the time, but it's a vague truism, not a theory.

>It's nice that they used to prance about, but now they're much quieter.

Huh?

I find all this talk about 'boumas' interesting. In fact, I do propose the theory that in most languages ( if not all) letter shapes and, to a lesser extent, word shapes are important from an entirely different perspective, that is, from a sign or iconic perspective rather than merely an arbitrary symbolic association. I describe this phenonemon in some detail on the first page ( home page) of my website, (Click on my name below.) With respect to boumas, the iconic vs. non-iconic details would still have to be worked out but I would certainly be a proponent of the view that iconicism plays a role in bouma recognition. But, of course, this would require too complex an analysis to describe on this thread, unless, of course, these threads were designed to encompass many pages and diagrams which I don't think they were. Finally, I take the position that immersive reading experiments do not rule out the importance of phonemic processing in word recognition since one can still consider phonemic processing on a fundamental, visceral level as mediating the actual word level recognition regardless of the time duration of the stimulus. It might be interesting to consider other experiments that might shed more light on the phonemic controversy but immersive reading experiments alone do not resolve the issue.

Dude, the independence of immersive reading from phonemics was proven before any of us were born. It's a given. Move on.

hhp

So was it at one time proven before any of us was born that the Earth was flat and prior to Galilieo being born that the Earth was the center of the Universe, (Geocentrism). Without intending to be offensive to anyone, I'm sticking to my position that immersive reading experiments do not prove phonemic independence, at least, not on a visceral level which could conceivably be the most important level.

>the independence of immersive reading from phonemics

Presumably because people who are born deaf are able to read.

How do they describe the process?
And how does this relate to their ability to speak?

Nick, that's a good point. But the reading and speaking proficiency of someone who is born deaf is substantially less than normal hearing individuals or someone who becomes deaf later in life. Sign language is the natural language of someone who is born deaf. And it can be said that learning sign language relies on the same kind of innate cognitive rules and information theory principles that learning written and spoken language depends upon. In the former, the rules are in the contextual framework of hand motions, finger movements and facial expressions, in the latter they are in the framework of sounds or symbols representing sounds. The fact that a deaf person can learn and speak written language, albeit, substantially less proficiently than a non-deaf person, verifies that with much diligence and effort written and spoken language might be learned on purely a symbolic associative level. But lets not make such an interpretation hastily. Though a deaf person cannot hear, they can still go through the motions of articulation and get a sense, on a rudimentary level, the iconic aspects of articulation. Furthermore, since the shapes of letters of the alphabet mimic the articulation associated with production of the phonemes represented by the letters, the deaf person can use letters as cues for associating the underlying phoneme. Thus, I would say even a born deaf person has some sense of phonemes by associating articulation with the articulation cues provided by letters. But how does this relate to the immersive reading experiments that rely on the data from non-deaf subjects? Perhaps, a good experiment would be to compare the immersive reading data of a non-deaf group with a born deaf group. Different results between the two groups might suggest that phonemes do play a role in immersive reading since the born deaf group's utilization of phonemic cues would be to a much lesser extent.

Immersive reading is free of the constraints of phonemics. This is old news.

hhp

Hrant, perhaps you'd like to point us to the proof that immersive reading independent of phonemics, by which I presume you mean completely and always independent of phonemics. It is not enough, these days, to say 'this is old news' or 'this was proven before any of us was born'. We want to know how it was proven. We want to pick at the experimental methodology of the studies that prove it. We want to see how well the proof stands up to counter-evidence from more recent studies.

I'm increasingly of the opinion that reading makes use of any and all information available, in a variety of different ways and, most alarming perhaps for anyone trying to determine 'the true model of reading', in ways that vary between individuals. The fact that deaf people can learn to read (although they typically read very poorly, because whatever their role in immersive reading phonemics are important in learning to read phonetically-based writing systems such as alphabets and syllaberies), does not prove that hearing people never use phonemic association during immersive reading.

Randomly pick up any decent book on reading theory - it's in there. There are no "recent studies" that disprove what's been known to be fact for many decades. Challenge me when it's a real gray area, but don't make us waste time proving that Man has landed on the moon.

Normal adults read "silently". Even if you can "hear" yourself saying the words, that's just a post-processing reaction. No mapping from shape to sound is necessary for reading, unless you're like 7.

> I'm increasingly of the opinion
> that reading makes use of any
> and all information available

Huh, I wonder where you got that. :-)

hhp

Here's an interesting article I found on the web. You might want to take a look at.
http://www.apa.org/monitor/mar00/dyslexia.html

One of the studies referenced by Kevin Larson during his ATypI presentation found that word recognition skill was affected by phonetic information in the following way: in sets of words with similar spelling, participants were slower to recognise words with variant pronunciation or incorrectly identified them. In other words, participants associated particular combinations of letters with particular phonemes, and faltered when the trying to recognise a word that contained the same letter combination corresponding to a different phoneme. This does not indicate a major rôle for phonetics in immersive reading, but it does indicate that phonetic information is not entirely absent or ignored.

I agree that 'no mapping from shape to sound is necessary for reading', but this is very different from asserting that such mapping never takes place for mature readers. I think we often run the risk of assuming that reading must be a very clean and efficient process -- indeed of assuming that thinking is a clean and efficient process --, but neural networks are not clean and efficient: they are profligate, and they can afford to be because of the sheer processing power that the brain possesses. Sure, you don't need to map from writing to phonemes in order to read, but that mapping already exists in the brains of everyone who has ever looked at words and figured out how to pronounce them.

Big giant Boumas ate little Hrant in his childhood nightmares.

Sorry, I'm stupid! I didn't resist... :)))

John, what you say makes sense, but recent fMRI studies of dyslexia subjects suggest something else, that the area of the brain involved in word recognition must be activated by another area of the brain involved in a phonemic learning phase. In dyslexia subjects, the word recognizer portion of the brain is not activated. This is consistent with the Oral Metaphor Construct,(OMC), which I describe on my website. The OMC capacity is genetically controlled and, thus, a defective gene could inhibit its functionality in the precise manner exhibited by dyslexia subjects. In dyslexia subjects, the hardwiring of phonemic-metaphoric mappings accomplished over long evolutionary epochs through a process I refer to as 'Knowledge Inheritance' is either absent or not routed to the visceral visual neural network resulting in the inability to activate the word recognizer portion of the brain. Thus, dyslexia subjects must consciously learn the orthographic phonemic metaphoric mappings to meaning which, of course, would be a slower and much more tedious task. Furthermore, it is my position that in normal reading subjects, the word recognizer portion of the brain, once activated, being that it involves unconscious hardwiring of phonemic-semantic relationships, would still be a phonemic based process utilized in what is referred to as 'immersive reading'. Confirming this is the fact that dyslexia subjects typically have no difficulty recogning diagrams, or images and, thus, do have a normal,non-phonemic,neural recognizer in place.

Hey there!
I'd just like to thank everyone who has posted a note on this website because it has helped me a lot with my studies!
Have an awesome day coz u guys really rock...