Archive through September 17, 2003

amills's picture

Acocdrnig to an elgnsih unviesitry sutdy, the oredr of letetrs in a wrod dosen't mttaer, the olny thnig thta's iopmrantt is that the frsit and lsat ltteer of eevry word is in the crcreot ptoision. The rset can be jmbueld and one is stlil able to raed the txet wiohtut dclftfuiiy.

Stuff to get you through the day. Brought to you by amills.

hrant's picture


Boumas in heavy action!


dyana's picture

That sounds like a very interesting study. Do you know where one could find it?

hawk's picture

wehre? wehn?

nad omst iomrtpant: #$? ywerf %%%=---,,,@@. Hweover, 8967--[gjhhytr km,,]= etete. kjghgj
13224222.789777. ? ro

Dan Weaver's picture

Some people have way to much time on their hands. Dan

hrant's picture

No Dan, this is actually serious! It's a [further] validation of Herman Bouma's research, which is very valuable for a text font designer. The outside envelope of frequent clusters of letters is much more important than the internal details of individual letters.

I too would like to see the original study in full.


hawk's picture

ohhh. btu i hvae antother stttudyyyyyyy. ot wrttite abobut iti

mitchell's picture

I read that quite easily. cool.

I saw a land rover advertisement kinda like this once. I think it replaced certain letters with a space instead of scrambling the words though.

eomine's picture

> this is actually serious!

It's hard to tell if you don't know who is doing it.
Is this really an academic study, Allison?

Anyway, it's interesting, indeed.

rs_donsata's picture

I could even read faster the scrambled word messages than i usually read english.

I demand a serious explanation !!!!

jcoltz's picture

Gordon Legge at the University of Minnesota has done some great work in the area of vision and reading psychophysics (especially with regard to low vision). He also happens to be a very nice guy (translation: I'm sure he'd love to talk boumas with you, Hrant)!

hrant's picture

Thanks for the great ref!


kentlew's picture

>The outside envelope of frequent clusters of letters is much more important than the internal details of individual letters.

Hrant, I don't see how this actually supports the "bouma" theory, as I have understood it. By rearranging the order of letters, especially in longer words, the envelope shape is significantly altered. The contour shape of "jmbueld" is very different from "jumbled". It seems to me that I am only reading it easily because I am deciphering the internal spaces and quickly rearranging.

Please explain how I am wrong here.

-- K.

John Hudson's picture

You're right, Kent, and the decipherment of jumbled words relies heavily on cnextot, especially if the the distribution of vewols and ctnnnsaoos breaks down the number of apparent syllables, and not at all on pttaoo.

John Hudson's picture

I think you're jumping to conclusions, Hrant, because another explanation for the study results is that we see collections of individual letters and know that those collections only occur in particular words, informed by context. So I see pttaoo, and rather than seeing a bouma shape defined by the p and o and some intervening rounds and a couple of pointy bits, I see a p, two t's, an a and two o's, and I know that collection of letters only occurs in potato. This is very much easier if the context is such that I'm expecting to see 'potato' in the text.

Note that I'm not saying that this is true -- I really don't know, any more than you do -- I'm just pointing out that there is more than one explanation for the test results as reported (we have not actually seen them yet).

hrant's picture

The basics (of which this subtopic is part) are really pretty simple once you have the model in your head.

> the olny thnig thta's iopmrantt is that the frsit and lsat ltteer of eevry word

This is for example is a severe over-simplification. Reading is not an all-or-nothing deal, it's heuristic. There are levels of certainty, and the more proficient the reader the tighter the thresholds are. That's for example why typos can be glossed over successfully. These levels arise from the interaction -at different priorities- of things like bouma envolopes, semantic context, individual letters, etc.

Certainly a word where all the middle letters are x-height is more easily read jumbled than one with a single "f" in it for example. BUT my point here is that the favoring of the first and last letters -irrespective of what's in the middle- supports my theory.

Part of my theory is that the external envelope (what I sometimes call silhouette - but I've realized that could be misleading because a silhouette can have holes while an envolope can't) of a bouma is more important than the insides. The insides are generally only used in a fallback mechanism. Bouma's findings* as well as this finding (assuming it does show what we think it's showing) support this aspect of my theory very well.

* See his papers: "Visual word recognition of three-letter words as derived from the recognition of the constituent letters", co-authored with Bouwhuis; and "Visual Interference in the Parafoveal recognition of initial and final letters".

Context: of course it's very important, but it clearly doesn't explain the favoring of the first and last letters. That's the clue here.


amills's picture

Eduardo-I don't know if it is an actual study. A friend sent this to me but I will do some research and if I find out anything I'll be sure to post! cheers.

eomine's picture

Hi people, just check this.
Dan was right.

William Berkson's picture

I don't know Bouma's theories, but these interesting examples I think refute the idea that the external profile of the whole word is the overriding factor in readability. The internal part of the profile is radically changed, and it is still readable.

I notice that in some cases the second letter being out of place also has a bigger affect on readability. It is interesting that the algorithm for spell checkers seem to act the same way; if you've got the first two letters right, they can guess the rest better than if everything else is right, and the first two messed up.

There is clearly some kind of mental heuristic system going on, as Hrant says, but it seems to me that the legibility of the internal parts of the first and last letters are likely an important part of the heuristic, and not simply the external profile.

The first and last letters being particularly important also means that making both the left and right halves of the letter easily distinguishable from other letters may be important to readability.

If no one has studied this phenomenon, I think there is a lot to learn from it.

hrant's picture

> The internal part of the profile is radically changed, and it is still readable.

Which of course means they're less important!
But you're right, never unimportant.

> I think there is a lot to learn from it.

Which is why I've been on it for ~5 years! :-)


hrant's picture

> I would hypothesize that it's similar.

No, heuristics is diametrically opposed to algorithmics.

> Context seems just as important

It's probably more important. But it's also outside the scope of typography, of what a type design or even typographer has control over.


William Berkson's picture

>The human brain may be more mysterious,
but I would hypothesize that it's similar.

I think that relatively simple algorithms will turn out to be part of the story, but only part. In a comparison, I don't think linguists have been able to characterize the syntax that every native English speaker knows intuitively, though they can get some of it.

The reading problem should be much simpler, I would think.

>No, heuristics is diametrically opposed to algorithmics.

I don't think this is accurate. George Polya, the Hungarian mathematician who pioneered heuristics and may have coined the term in this sense, had heuristics as a series of structured questions. The difference with algorithms is that they normally come to a definite conclusion (loops are possible, as in programming failures.)

Polya's 'How to Solve It' is the source.

I suspect that if and when the science is understood, there will be a pretty wide latitude for personal expression with little difference in readiblity.

What is interesting about the scrambling is that it does seem to be a way to test hypotheses about the algorithms for disciphering words. Might help understand dylexia.

hrant's picture

"And so to completely analyze what we do when we read would almost be the acme of a psychologist's achievements, for it would be to describe very many of the most intricate workings of the human mind, as well as to unravel the tangled story of the most remarkable specific performance that civilization has learned in all its history."
- E B Huey

Reading will never be completely understood, because of
a simple tautology: a system can never understand itself.

> heuristics as a series of structured questions

I think heuristics is based on not getting definite answers - the opposite of what an algorithm is about. Human reading is all about mushy, dynamic thresholds, it's about making the right amount of mistakes. That's why an algorithm will never fail to get tripped up by typos, while a good human reader will assimilate and gloss over them.

> Might help understand dylexia.

A lot of effort goes into understanding and treating "exceptions" such as dyslexia. It's a shame, when we are still so far -and in some ways getting further- from understanding the main rules! This is a textbook failing of contemporary Western science. The Romans threw away all the masterful subtlety the Greeks had so nicely compiled form the truly ancient cultures.


paul's picture

I've mentioned this before, a while back, but one of the most fascinating books I

timd's picture
This guy has written a script to achieve this.
a definition of copyright

William Berkson's picture

>heuristics is based on not getting definite answers

I think you are misunderstanding heuristics, at least as conceived by Polya. Heuristics are a series of questions that are guidelines to problem-solving. The answers can be very definite. The difference is that there is a creative gap between question and answer, which can involve a lot of trial and error. In an algorithm, it is a completely set and pre-defined process.

>a textbook failing of contemporary Western science.

Hrant, You misunderstand the nature of scientific method. The 'exceptions' are key tests of general theories, which must cover all cases. Jonathan Swift, author of Gullivers Travels, similarly said that 'natural philosophy' should be called 'unnatural philosophy' because it dealt with such weird stuff. He didn't understand it either.

Weird stuff such as electricity and magnetism proved a key to understanding universal principles, and to such technological breakthroughs as the computers and networks we are communicating through.

In the case of reading, understanding dyslexia as a test case will, I suspect, be a key to understanding the general normal processes.

Incidentally, G Noordzij in 'Letterletter' claims that he was able to cure dyslexic children by training them to be aware of the difference in white spaces between the letters.

If you want to understand scientific method better, I recommend 'Conjectures and Refutations' by Karl Popper (under whom I did my PhD in history and philosophy of science.) For one piece of the electricity and magnetism story you can read my own 'Fields of Force: The Development of A World View from Faraday to Einstein.' --Set in van Krimpen's Spectrum!

hrant's picture

Paul: very interesting.

"Heuristics": inventing a word is great, but that's not what it means now.

Dyslexia as an exception: of course the Scientific Method (which is these days more of an excuse to make more money and not a method at all) involves analyzing exceptions. The problem is that contemporary science/academia doesn't care for that: it's become a way to milk the exceptions for their own sake. The Scientific Method has become an end, instead of a tool. The true human method is that of the ancients.

Noordzij: he didn't cure jack.


William Berkson's picture

>Greeks had so nicely compiled form the truly ancient cultures.

Ok, this is off topic. The Greeks did not simply 'compile' from ancient cultures. They criticised more ancient cultures and created new ideas never seen before in history. Archimedes is an outstanding example here. Similarly in modern science, which took its point of departure from Greek and Islamic science, it criticised the past and came up with new, superior theories.

hrant's picture

Like depleted-uranium cluster bombs.


ponofob's picture

I only know heuristic as an adjective but in my book (which is semiotic), a heuristic hypothesis, for example, is an hypothesis which drive to a discovery procedure; an hypothesis which isn't true, or false, but which product a discovery procedure, which can be another hypothesis etc. A discovery procedure being the way of something is analysed, at least in semiotics. But both my english unease and the fact i know few let me think what i say may be distorded.

William Berkson's picture

The word 'heuristic' according to the dictionaries I have checked comes from the Greek 'heurisko' (sorry for the roman letters) meaning 'find'. Anything that is an aid to discovery can be called heuristic. Heuristic procedures may lead to definite answers, or not, as the case may be. For example, in the computer context it means, according to the Concise Oxford Dictionary (1990 edition) "proceeding to a solution by trial and error."

Thus Hrant's understanding that "heuristics is based on not getting definite answers" is mistaken.

Guillaume's usage is also accurate in English.

Another dictionary dates the usage of the word(in German) to 1821.

William Berkson's picture

The procedures inbedded in the brain, some of which are described in the book noted by Paul, are evidently heuristic in the correct sense. Normally they lead to a rapid and definite, though sometimes mistaken diciphering of a word. The procedures can also fail to come to a definite conclusion also, as when we fail to dicipher a piece of handwriting.

Both the mistakes of normal and dyslexic readers I think would be important clues to what the heuristic rules are that are imbedded in our brain. On the positive side, our ability to read mixed up words, as noted in this thread, is also an important clue.

hrant's picture

A term is only useful to the extent it represents something in reality accurately. The way William is using the term might be what the inventor of the word intended, it might even be the way a dictionary defines it (although dictionaries are very good at destroying subtlety) but to me it's simply not the way the brain works. The way the human brain works (as far I as believe - but that's always an automatic disclaimer for everybody at any time) is fundamentally different than the way a computer algorithm works in that it doesn't care to arrive at a "correct" answer - it only cares to move on to the next problem with sufficient certainty. Controlling the type and amount of mistakes is integral to this, whereas an algorithm cannot accomodate mistakes. It's really a very simple distinction, but a lot of people have a problem with it because it undermines the entire premise behind western science. If you think a human thinks algorithmically like a computer, you need some serious debriefing.

My problem with western science -the way I see it as starting from Rome, as a qualitative if subtle break from Greece- is simply that it's too inward-looking, too concerned with the algorithmic results. As a result, it will only go for example from an understanding of dyslexia to one of normal reading by circumstance, not naturally. This is -again- the opposite of the what I might tentatively call the Ancient Way. So the problem with the western study of "exceptions" is that -in practice- the whole point is as a rule made secondary. Just like TV: theoretically very useful to culture, but in practice a tool for brainwashing and making the rich richer. Why do you think indifferent philanthropy is now gone, replaced by large corporations funding university research? Because it's better for the economy. The whole system is decrepit.

But to be fair, I guess I had a too-strong reaction to William's "superior". The West is not superior, not by a long shot. It's not how far you can launch a satellite, it's what you do with it. The West is in fact destroying the world; converting culture to cash, at increasing low rates.


William Berkson's picture

>it doesn't care to arrive at a "correct" answer

From an evolutionary point of view, the brain must have an answer that is correct enough - corresponding to reality - for the species to thrive and breed. Correctness to that extent counts. Literacy has recently also become an economic survival skill.

I don't think Hrant's comments on science are well informed or correct. But to critique them would take this too far afield.

Dan Weaver's picture

I would only ask you to think about the distruction of language if misspelling was expected as the norm. My example is about eubonics (and this is probably misspelled). There was a eubonics contest for Miss America all states were represented except I da ho. In the late 19th century if you misspelled a word you got a ruler rapped across your knuckles and you never misspelled that word again. So is it allowable for us to be lazy and anticipate the next word, or should we really be reading.

hrant's picture

I'll avoid getting into the pros and cons of corporal punishment...

> is it allowable for us to be lazy

It's not up to us: our subconscious is inherently "lazy", although in the best possible meaning of the word.


matha_standun's picture

From an evolutionary point of view, the brain must have an answer that is correct enough - corresponding to reality - for the species to thrive and breed.

Evolution, brain and reality reunited at last. The postmodernists will be pulling their hair out.
What we call Reality surely gets into the brain via language - a very slippery thing altogether - so correct is maybe not a good word for the answer it arrives at, unless we're willing to admit that correctness is relative.

Is that really thriving that we're doing?

Hey, I've just seen HHP and Dean Swift in the same paragraph.
Fame at last for the Dean!!


William Berkson's picture

>Reality surely gets into the brain via language

No, perception of objects, eye-body movement-object coordination, etc is first.

Understanding of language comes afterwards, based on the above. Correct is not relative. You walk into the door and get a black eye, or you walk through the doorway and are laughing. Reality has a way of punching us in the face when we try to ignore it.

>Is that really thriving that we're doing?

Evolutionarily, yes, there are billions of us. Morally, probably most would agree with Swift...

matha_standun's picture

Reality cannot be seen.

It has nothing to do with eye-body movement-object coordination. 'Reality' is a word for our interpretation of what is going on around us. It is by definition a relative concept.

Correctness exists inside the framework of reality.


matha_standun's picture

>'Reality' is a word for our interpretation of
> what is going on around us

This can only be achieved through language.
Language is slippery.


John Hudson's picture

I don't use 'reality' for my interpretation of what's going on around me: I use 'interpretation' for that. Reality is beyond me, in more ways than one.

William Berkson's picture

>off the hook if a typo gets through

As noted above by Frederik, spelling and proof reading are also not my strong points. I certainly don't pretend it is anything other than a failing. A few typos in a book is not a cardinal sin, though.

>Reality is beyond me, in more ways than one.

I am with John on this. The concept of reality is something beyond us, which we try to grasp as best we can. We always have an imperfect grasp of it. Philosophical idealists say we can dispense with the concept of reality.

The idealists are wrong, because our efforts to understand the reality outside ourselves, including other people, is fundamental to our humanity, as well as to science.

Philosophical idealism, as Jorge Luis Borges pointed out, is an excuse for wallowing in vanity. It is also a dangerous excuse to unleash unreason upon society. It poisons discourse and politics. Large parts of academia are sickened by it.

ponofob's picture

i'm astonished about how some people are so sure of what they say or think they know. Matters like how we analyze things first, if it's language or perfection first etc. are such uncertain things, so unprovable in anyway that it appears to me a little pretentious to say : i have the truth, it's this, or that.
There's an interesting idea about how we first aprehend things, the idea of no

matha_standun's picture

I don't use 'reality' for my interpretation of what's going on around me: I use 'interpretation' for that. Reality is beyond me, in more ways than one.

What do you use the word 'reality' for then, John? Would you be willing to agree that there are enormous differences of opinion about what it signifies? or that it's meaning is not fixed?

matha_standun's picture

What reality 'is' depends on what language you are speaking, your culture, your religion. How we interact with what it is is relative, how we prove what is there is there is faulty and reflects thelimitations of our tools and languages and brains.

We can live without proof though. That's where faith comes in.


William Berkson's picture

>What reality 'is' depends on what language you are speaking

No, the whole notion of reality is that it is not language dependent. Our certainty about what is real is another matter. We are never sure we are right, and in fact we often err, but that is another matter.

If I say 'I have three children' or 'J'ai trois enfants' or 'Ich habe drei Kinder' or 'Wo you san haizi' or 'Yesh li shalosha banim' it doesn't change the reality. I do have three children. I may of course be wrong. There may be an extra kid I don't know about from an old girl friend, but the reality is what it is.

And when a person swears before a court that he or she promises to 'tell the whole truth and nothing but the truth' we know what is meant - even if there is a translator involved.

The whole effort to deny these obvious facts, a move influenced by Husserl, is destructive. It is Husserl in fact who wanted certainty, and thought he could get to it by only describing perception, and 'bracketed' the question of whether it corresponded to reality.

The rottenness of this move becomes evident when we consider the SEARCH for truth, whether it is in science, the courtroom or politics. There truth - correspondence to reality - becomes a regulative idea (as noted by Kant) without which, for example, contradictions are not a problem.

It is this abandonment of the critical tradition (a tradition started especially by Socrates) which is the poison in philosophical relativism and idealism. Heidigger's support of Naziism and Sartre's support of Stalin are indicators of the evils that relativism can be used to support - because, having denied one reality, evidence doesn't matter, and you are left with only irrational commitment, and one commitment is as good as another.

The fact that relativism is fashionable in the humanities, particularly on the European continent, does not make it any less odious or misguided.

I don't think this has a lot to do with type, but since it has repeatedly come up on this board (I avoided comment before), I have commented.

ponofob's picture

when you say "i have 3 children" what do you mean exactly ? That you own them ? In a context it can mean you have them in a factory where they make nike shoes; in another that you're working in a school and have a very little class

hawk's picture

hey....something to relax. and something about truth.... :-)com

hawk's picture

hey....something to relax. and something about truth.... :-)

hrant's picture

William, it is your narrow-mindedness that's destructive. Your attitude has caused much more misery in this world than relativism.


matha_standun's picture

relativism is a way of thinking with doubt

and accepting that sometimes different things can be correct or true or right at the same time.

This is not a license to do whatever you like, William, and linking relativism with nazism is a cheap ploy. It's been done before and the only thing it has going for it is that it covers up a weak argument very nicely. The nazis used words like 'science' and 'truth' a hell of a lot while they were up to their necks in blood.

And as for the search for truth, we build our own analytical tools inside the system we are trying to analyse, invent our own notation system to work whatever results we manage to glean into something coherent in our own eyes and we call the result 'reality'.

Truth in a courtroom or in politics? That's exactly where you'll find it, defined in a way that suits those in power (the ones that pay the scientists).

Try and search for truth without language and see how far you get, William.

Syndicate content Syndicate content