Somebody Agrees with Hrant about Black-Letter

Jasper de Waard

I will look up some research on this when I have the time. In the mean time:

@Thomas Phinneys point is important here. Saccade length indeed refers to distance (which has little effect on time), not time. Fixation duration is another matter.

Furthermore I think it is important to note that covert attention (directing attention to a location without moving our eyes there) also plays a significant role in reading. Much can be deduced from eye-tracking studies, but we still don't really know whether covert attention moves stepwise (like saccades), taking in chunks of letters simultaneously, or more smoothly (approaching a letter-by-letter approach).

Reading is a layered process, starting with visual processing, then linguistic processing, and then feedback loops to create and adapt expectations of what's to come. I think the visual processing is usually faster than the linguistic processing (these are quite seperate brain regions), so that the gains of a 'better' writing system would be minimal.

Furthermore, I really see no way of reliably testing whether some writing systems are better than others. @Hrant H. Papazian you seem to have ideas, I'm curious.

John Savard

"Better" is a loaded word. But if I recieved news that speed readers in Korea were accomplishing feats beyond the reach of the fastest speed readers in the United States, I'd consider that at least a hint.

After all, a laboratory study only measures the performance of its test subjects, and reading ability is variable between individuals. So I think knowledge is attainable, but perhaps a lab is not the best place to obtain it.

John Savard

Jasper de Waard said:

 I think the visual processing is usually faster than the linguistic processing (these are quite seperate brain regions), so that the gains of a 'better' writing system would be minimal.

In that case, one possibility is that the sole beneficiaries from an improved writing system might be geniuses who happen to have unusually fast linguistic processing. I suppose this could be a background detail in a science-fiction story.

In any case, this image

shows what the sort of "improved" writing system I was thinking of, inspired by Hrant's comments, would look like. Unlike Hangulatin, although it might work with a lot of existing typefaces, and it certainly gives the words lots of bouma... I have to admit it's quite ugly in appearance, and thus I can't really recommend such a sacrifice of beauty to a hope of utility.

Hrant Հրանդ Փափազեան Papazian

Thomas Phinney said:

If you keep linguistic-content per saccade fixed across writing systems, and the spatial part is not critical (being below the threshold of what people can take in visually), then time-per-saccade could still be a variable.

I'm not sure (and I think you mean fixation duration, because saccade duration is always near-instantaneous). I think fixation duration either doesn't vary much, or it's proportional to saccade length. Which is a relief because we can focus on much more easily measured things like saccade length and especially net reading speed (assuming comprehension).

I believe that shorter saccades –both between and within writing systems– are generally an indication of non-immersion; this happens in English too (which is why the deliberative reading speeds typical of lab tests are so slow). In deliberative reading, the writing system, just like the font, is moot (unless it's really horrible).

BTW choosing to believe things might seem facile, but because exactly as you say robust testing of this is so extremely difficult, contemplation of the little we do know is all we have (if also a bit of gut feeling). Managing uncertainty is integral to decision-making... (And some people get riled when others are too cavalier for their tastes.)

Jasper de Waard said:

Furthermore I think it is important to note that covert attention (directing attention to a location without moving our eyes there) also plays a significant role in reading.

I just call that reading the parafovea during a fixation. This is where immersion kicks in (including much longer saccades) and boumas come out to play. Poor readers, as well as good readers who are not motivated, don't benefit. But we should still design for it, not least because it doesn't harm anything. Well, except the life of the designer.  :-)

Jasper de Waard said:

Furthermore, I really see no way of reliably testing whether some writing systems are better than others. @Hrant H. Papazian you seem to have ideas, I'm curious.

It's very difficult. Even simply testing fonts in English has proven an extreme challenge!

I think the key, and the hardest thing to pull off, is preventing the test subjects from being self-conscious of being tested... :-/  In line with @John Savard 's "perhaps a lab is not the best place to obtain [knowledge]."

But the lab can be clever. One could obtain permission from subjects to observe them "in the wild", and collect data furtively. For example a cafeteria could be fitted with cameras. We might have the technology already.

The good news is I don't think you need a large number of test subjects, because presumably the human brain is pretty consistent, You do need a few very good ones however.

Hrant Հրանդ Փափազեան Papazian

John Savard said:
one possibility is that the sole beneficiaries from an improved writing system might be geniuses who happen to have unusually fast linguistic processing.

Actually any –proficient– reader faced with easy text benefits. Which is probably most of the reading we do, and to me the only reason laymen would report serif being easier to read than sans for example.

John Savard said:

In any case, this image

shows what the sort of "improved" writing system I was thinking of

That's pretty much what I messed with in 1999. See the last thing here:
http://typedrawers.com/discussion/comment/39625/#Comment_39625

Russell McGorman

Does more data per unit of foveal field translate to faster reading, or more time spent interpreting data per saccade? 

Thomas Phinney

@Russell_McGorman
I expect neither, but shorter saccade distances instead. As pointed out previously in this thread, we know from research on other languages that saccade distances are shorter for readers of ideographs, who are taking in more content per character and have more information-dense texts.

This is why I do not expect that John Savard’s improved writing system would do what he hopes. It is solving a problem that I suspect we do not have: I suspect saccade distances for western languages are based on cognitive limits of information density, rather than limits of human optics. This appears to me to clearly be the case for ideographs.

John Savard

Incidentally, I think that the "rounded gothic", not referred to as an early Roman by anyone I know, used by Gutenberg in the Catholicon is perfectly legible, so this shows some Gothic types are legible even by modern standards.

Hrant Հրանդ Փափազեան Papazian

Thomas Phinney said:

It is solving a problem that I suspect we do not have

It's irrational to believe that a writing system cannot be improved.

Hrant Հրանդ Փափազեան Papazian

Russell_McGorman said:

Does more data per unit of foveal field translate to faster reading, or more time spent interpreting data per saccade? 

The simpler the content (and most content is simple enough for most people) the more the mechanics of reading becomes overheard; so more data per fixation (notably including the parafovea = immersive reading) is indeed a good thing. This should be self-evident. Imagine if English grammar required a blank space between each letter (and let's say a bigger space between words); since saccades cannot extend beyond an area that has not yet been observed (and acuity drops off precipitously) it's obvious reading would slow down.

Latin (and Armenian FWIW) wastes vertical acuity.

Thomas Phinney

The evidence is to the contrary.

In Chinese/Japanese with han characters, information density is higher than in western languages. Readers’ saccades are proportionately more tightly spaced, and the same linguistic quantity of info is absorbed per saccade as with readers of English and other languages.

Again: this strongly suggests that the currently-seen limit of info-per-saccade is not a visual limit, but a cognitive one.

I think it is totally reasonable to assume that at some level there would be visual limits on info-per-saccade. It's just that with our existing languages and writing systems, those are not the limits we normally hit. We know reading speeds remain reasonably constant for text within a significant range of sizes, and slow down for super-large (or super-small) text. It may be that the reason for the large-text slow-down is indeed visual limits for saccade size—seems like a fair hypothesis. (Or it might be that time to get from one saccade to the next goes up. I would not be surprised if the answers are already well known to cognitive psychologists.)

Hrant Հրանդ Փափազեան Papazian

It makes no sense to believe that the visual representation of language does not impact reading performance. It makes much more sense to believe that the testing has been insufficiently sophisticated to reveal the performance role of visual representation.

Thomas Phinney

I said no such thing; that's a straw-man caricature of my argument.

Robin Mientjes

Hrant, you clearly want to see different testing done to find answers to what we don’t know – but people are reporting on what we do know. This is a good time to step back, stop putting words into people’s mouths, and maybe consider proposing what sort of testing can be done, or who might be appropriate experts, or who could be a partner in this sort of testing.

Thomas Phinney

Every time testing does not show the results Hrant believes in (which is frequent), he just says it must not have been done correctly, it must not have been “immersive” reading.

Hrant Հրանդ Փափազեան Papazian

Robin Mientjes said:

Hrant, you clearly want to see different testing done to find answers to what we don’t know – but people are reporting on what we do know.

All of which I've done before.*
But I'm not waiting for Capitalism to sanction my views to try to help my users as best I can, versus using flawed testing (ergo: what we want to believe we know) as an escape.

* Reading matters:
http://typedrawers.com/discussion/comment/39808/#Comment_39808

BTW following people's train of thought to their [il]logical conclusion is not putting words in their mouth; it's a discoursive service (because a public conversation is not about its participants).

Thomas Phinney

My prior post had content which was in fact the exact opposite of what you, Hrant, claimed I was saying.

I am done with this discussion.