Notes on Navigation 1

One of the problems that fascinates me is the problem of navigation. Not web navigation, although that is interesting in its own right, but navigation in the physical world: how we get from one place to another, whether its on a route we already know, or a new route, how we make use of instructions and artifacts like maps. (Edwin Hutchins has, in Cognition in the Wild, looked at how groups of people collectively navigate at sea.)

Like Lucy Suchman, I don't think that to go from one place to another, human beings simply make plans in their heads and execute them. I think navigation is a complex activity, where many things come together: the plan that a person makes, the artifacts he uses like maps, written instructions as well as other artifacts like road signs all of them shape the experience in complex ways.

I had an interesting experience on a recent trip to Spain that is related tangentially to navigation. Now previously, when I went on trips abroad, I read a page or two of Wikipedia to find out what the noteworthy sites were. A best, I would write down their names, at worst, I would just try and remember the names in my head. The trips didn't turn out badly and I've always enjoyed just walking randomly on the streets of a new city without any particular destination in mind. But I don't remember much of these trips either. One of the reasons for this could be that I did not annotate the photographs I took immediately after the trip while the memories were still fresh. Another reason could be that I did not keep a journal of the trip. But I think there may be a third reason as well.

When I visited Amsterdam with a friend of mine, I noticed that he did things differently. He used a guidebook, a good one and usually planned to visit a couple of spots a day. No explicit planning was done, at least not in much detail but we always had the guidebook in hand as we walked around. We were always trying to map what we saw to what the guidebook said (well, he did the hard work but he explained it to me once he'd gotten it anyway). Just having the guidebook in hand, consulting it, somehow made the excursions much more informative. And more importantly, I remember much more of the trip! Even though, like the previous ones, I did not take photographs or keep a journal.

My hypothesis is that it was carrying the guidebook around that made a difference. Because we were always cross-referencing the guidebook, because our daily outings consisted of visiting a couple of locations in the guidebook, because even when we were just randomly walking the streets, our activity was directed: all of it had the effect of etching the details of the trip more firmly in my mind.

I got a chance to test this hypothesis when I recently went to Madrid for the WWW 09 conference. This time, I made sure that I did things the way my friend did in Amsterdam. I bought a guidebook for Spain: the National Geographic. I didn't plan on visiting too many spots on the same day, just one or two.

And you know what -- it worked! My trip seemed more substantial and I seem to be able to recall it better! Here are some observations on why and how having a guidebook made a difference:

• Reading the guidebook before the trip in detail proved to be unnecessary. I had time to just skim it on the flight to Spain and that was sufficient.
• It was much more important to have the guidebook with me as I was sightseeing: having it in hand, where I could refer to it easily, rather than keeping it, say, in my knapsack.
• I used the maps in the guidebook, especially on the day I took the Plaza Mayor Walk that was recommended. But the guidebook also helped me navigate the museums! The following point talks about how it did this:
  
• For the museum each exhibit is equally important. So one tends to get lost in a museum, getting overwhelmed by the sheer amount of stuff on display. The guidebook, on the other hand, will tell you what the important pieces in the museum are and more importantly, how to get to them (mine did, and I think a good one should!).
  
  • This meant that my activity as I went around the museum was directed: I was no longer just walking around but walking around with a purpose. I would continually refer to my guidebook to see where I was, both physically, and in terms of paintings/exhibits that I had already seen and the ones I still had to see. It was the same when I was walking on the road: my activity was directed towards the next landmark I wanted to see. Even if I was walking around randomly, it was still around a certain landmark.
    
  • At some point (this was in the Prado) my guidebook proved insufficient for the spatial navigation i.e. the instructions it gave me for moving from one room to another were outdated since the paintings had been moved around (obviously). So I picked up a floor plan, which had a map of the rooms with the painter(s) each room was dedicated to. I used the guidebook to find the next painting I wanted to see, the plan to figure out where I was and where I wanted to go, all the while looking at the paintings I was surrounded by to figure out which room I was in. -- it seems complicated to explain it but moving back-and-forth between these things proved surprisingly easy. I think a similar process takes place when we are trying to find our way in an unfamiliar place in our day-to-day life.
    

So where is this leading? Nowhere, really. And are there any takeaway points?

• One comes to mind: guidebooks should be designed keeping in mind that the best way for a person to use them is in-situ, in the process of sightseeing. And since my guidebook served the purpose so well, I am guessing they already are designed that way.
  
• For me at least, sightseeing became much more productive when it was directed than when it was random. I remember much more from this trip than I do from all of my others combined. It was because I had the guidebook with me so I had a name to go for every street I walked on and perhaps names help you remember better? Or to turn the argument slightly around, when I look at the guidebook today, after the trip, the images in my mind combine with the words on the page and the memory gets solidified. Or something. Sorry for the mixed metaphors but I haven't found a technical way of saying what I want to say.

That's all! More points about navigation to come later.

Calling all scientists and mathematicians

[This post doesn't really fall into the topics this blog is supposed to be about, but it's reasonably close. And besides, my other blog is too frivolous for it anyway.]

Jerry Coyne has a new essay in the New Republic on the split between science and religion, about "the never-ending attempt to reconcile science and religion, and why it is doomed to fail". Blah blah. I am of the firm opinion that the matter cannot be settled either way and this essay, like most others, manages to settle nothing. (See this rather interesting, but rather tendentious, response by Jim Manzi and this one by Alan Jacobs).

The reason for this post though is that I found some of the reasoning that Coyne uses near the end of his essay to be grossly misleading, although again, I can't make a cut-and-dried case for anything. Here is the objectionable graf:

The NOMA [Stephen Jay Gould's idea of science and religion being Non-Overlapping MagisteriA] solution falls apart for other reasons. Despite Gould's claims to the contrary, supernatural phenomena are not completely beyond the realm of science. All scientists can think of certain observations that would convince them of the existence of God or supernatural forces. In a letter to the American biologist Asa Gray, Darwin noted:

Your question what would convince me of Design is a poser. If I saw an angel come down to teach us good, and I was convinced from others seeing him that I was not mad, I should believe in design. If I could be convinced thoroughly that life and mind was in an unknown way a function of other imponderable force, I should be convinced. If man was made of brass or iron and no way connected with any other organism which had ever lived, I should perhaps be convinced. But this is childish writing.

Similarly, if a nine-hundred-foot-tall Jesus appeared to the residents of New York City, as he supposedly did to the evangelist Oral Roberts in Oklahoma, and this apparition were convincingly documented, most scientists would fall on their knees with hosannas.

Scientists do indeed rely on materialistic explanations of nature, but it is important to understand that this is not an a priori philosophical commitment. It is, rather, the best research strategy that has evolved from our long-standing experience with nature. There was a time when God was a part of science. Newton thought that his research on physics helped clarify God's celestial plan. So did Linnaeus, the Swedish botanist who devised our current scheme for organizing species. But over centuries of research we have learned that the idea "God did it" has never advanced our understanding of nature an iota, and that is why we abandoned it. In the early 1800s, the French mathematician Laplace presented Napoleon with a copy of his great five-volume work on the solar system, the Mechanique Celeste. Aware that the books contained no mention of God, Napoleon taunted him, "Monsieur Laplace, they tell me you have written this large book on the system of the universe, and have never even mentioned its Creator." Laplace answered, famously and brusquely: "Je n'avais pas besoin de cette hypothese-la," "I have had no need of that hypothesis." And scientists have not needed it since.

In a common error, Giberson confuses the strategic materialism of science with an absolute commitment to a philosophy of materialism. He claims that "if the face of Jesus appeared on Mount Rushmore with God's name signed underneath, geologists would still have to explain this curious phenomenon as an improbable byproduct of erosion and tectonics." Nonsense. There are so many phenomena that would raise the specter of God or other supernatural forces: faith healers could restore lost vision, the cancers of only good people could go into remission, the dead could return to life, we could find meaningful DNA sequences that could have been placed in our genome only by an intelligent agent, angels could appear in the sky. The fact that no such things have ever been scientifically documented gives us added confidence that we are right to stick with natural explanations for nature. And it explains why so many scientists, who have learned to disregard God as an explanation, have also discarded him as a possibility.

There is an interesting sleight-of-hand at work here. "If a nine-hundred-foot-tall Jesus appeared to the residents of New York City", Coyne says and if this was "convincingly documented", most scientists would start believing in God. You think so? I don't. There are a number of reasons why I think so but they call center on the meaning of the term "convincing documentation". How does one decide whether the documentation is "convincing"? Is there a definition of the term that most scientists would agree on? Because you see, if the documentation is not convincing enough then I can reject the nine-hundred feet tall statue of Jesus as the work of a scam artist while still asserting that science is not a priori indisposed to the idea of God.

Coyne further goes on to say that science is not philosophically committed to materialism; that rather, this is a pragmatic commitment, that scientists use it simply because it works. Again, I am unconvinced. I agree with Giberson's claim that "if the face of Jesus appeared on Mount Rushmore with God's name signed underneath, geologists would still have to explain this curious phenomenon as an improbable byproduct of erosion and tectonics." Coyne says this is nonsense, that were this to happen, he and his fellow scientists would convert to Christianity and take the scriptures literally. He, in fact, goes on to list other phenomena that will make him believe in God: "faith healers could restore lost vision, the cancers of only good people could go into remission, the dead could return to life, we could find meaningful DNA sequences that could have been placed in our genome only by an intelligent agent, angels could appear in the sky".

Nice try. But I'm not buying. My contention is this: should any of these things happen, not one scientist will take them as evidence of the existence of a Christian God. Rather, several other things will crop up: the validity of the documentation will come up, fraud will be suspected -- and at the end of it all, most scientists will believe that there is some unknown causal explanation at work here, something that will surface once more work is done -- and get to work on it pronto. (What do you think?)

Why does Coyne bring this up? I suspect this is a rhetorical gambit intended to play upon the image of the man of science in our culture today. This image, which most of us have, says that the scientist is a man who changes his mind by looking at the facts, that he never lets his facts be dictated by theory, that instead he changes his theories once the facts change. As opposed to this, the religious believer is one who doesn't change his mind even when when the facts say otherwise. So Coyne is saying: "Look, we scientists are ready to change our minds if such and such happens, whereas religious believers aren't even prepared to say such a thing!"

I am not saying that there isn't some truth to this popular image of religion and science. And I am pretty clear about which side I am on. But this notion in the popular imagination that there is something called a "fact" which scientists have easy access to, and which they use to change their theories has got to be contested -- at least a little bit. One of the reasons why science is so successful at what it does is precisely because scientists do not change their minds easily, that they stick with a problem, despite many setbacks they encounter at solving it. Thomas Kuhn's point in The Structure of Scientific Revolutions (a book I will talk about in length perhaps in another blog post in the future) is that the development of science proceeds in two alternate steps: a conceptual revolution that is followed by a period of what Kuhn calls "normal science" which is again followed by a conceptual revolution. In "normal science" scientists settle down with problems, which they model as puzzles. In doing so, they build their concepts further. While doing "normal science", scientists will massage their concepts to fit the data but they refrain from changing their concepts drastically. In other words, they do not lose their faith in their concepts, do not abandon them until a "crisis" (Kuhn's term) is reached. Then, and only then, do the concepts get radically revised.

I am oversimplifying what Kuhn says. But my broader point is that scientists have to have some kind of "faith" in their concepts in order to do normal science -- no science would get done if concepts were radically revised at the drop of a hat! The terms "normal science" or "puzzle solving" aren't meant to be pejorative either. They may sound unexciting -- who, after all, would want to do "normal" science? But the point is that these puzzles themselves are extraordinarily challenging even for the best and the brightest -- and cracking them gives a scientist fame and recognition among his colleagues. The community of scientists, to quote a metahpor used by (I think) Michael Polanyi, is like the blind man with a cane. He uses his cane to see what comes ahead, he gets better and better at using it with practice but he also has faith that the ground will not drop away under his feet suddenly. Scientists grope their way to knowledge. My point here is that science, like liberalism, has a set of implicit metaphysical commitments. These are not unchanging, they have evolved over the last five hundred years but nevertheless they exist.

Here's a thought experiment to see what I mean. Suppose there is an omnipotent God. Suppose that there is a woman in the last stages of a deadly cancer. Finally suppose that there is a faith healer. The faith healer appeals to God to save the woman. God decides that he will grant the healer's wish and the woman's cancer starts going into remission. Will the scientist accept the healer's explanation that God did this? Well, no. To a scientist a result is only a result if it is replicable as widely as possible. So the said scientist takes the healer to another woman who has the same condition as the healed woman, more or less. Now however God decides not to honor the faith healer's wish; there is no remission. Ergo, the scientist's point, that the faith healer is a scam artist, and the woman's remission was due to some unexplained causes which further research will bring to light. Well, you say, what will it take for the scientist to accept the healer's point, that it was indeed God who healed the woman? Here's my guess. He would take the healer to N number of similarly ill women and ask him to heal them. Even if all of them did get better (God allowing), the scientist is still free to think that the reason for this is some yet to be explained metabolic similarity between these women. Indeed, what may change is that the scientist now believes the healer to be sincere yet self-deluded and not, as he believed earlier, a scam artist.

I am not really sure where I'm going with all this. And I am sure there are many loopholes in what I have said. But that's just my point: there is no water tight way of saying that science is incompatible with literal belief. And frankly, aside from the policy implications for a democracy, by itself, it is an uninteresting question by itself.

From Edge, a talk on consciousness by the Gibsonian philosopher Alva Noe. I particularly liked this metaphor he uses:

A much better image is that of the dancer. A dancer is locked into an environment, responsive to music, responsive to a partner. The idea that the dance is a state of us, inside of us, or something that happens in us is crazy. Our ability to dance depends on all sorts of things going on inside of us, but that we are dancing is fundamentally an attunement to the world around us.

Why you shouldn't trust articles that say "Brain scans have proved this .. or that"

Great piece from Scientific American that cautions about using the results of brain scan experiments to deduce too much. My favorite part?

I visited neuroscientist Russell Poldrack’s laboratory at the University of California, Los Angeles, and arranged to get my brain scanned inside its MRI machine. Scanners typically weigh around 12 tons and cost about $2.5 million (not including installation, training and maintenance, which can drive the typical bill up by another $1 million). Right off the bat I realized how unnatural an environment it is inside that coffinesque tube. In fact, I had to bail out of the experiment before it even started. I had suddenly developed claustrophobia, a problem I had never experienced earlier. I’m not alone. Poldrack says that as many as 20 percent of subjects are similarly affected. Because not everyone can remain relatively relaxed while squeezed inside the tube, fMRI studies are afflicted with a selection bias; the subject sample cannot be completely random, so it cannot be said to represent all brains fairly.

A person jammed into the narrow tube also has his or her head locked firmly in place with foam wedges inside the head coil—nicknamed “the cage”—to reduce head motion (which can blur the images) before the experiment begins. The MRI scanner snaps a picture of the brain every two seconds while the subject watches images or makes choices (by pushing buttons on a keypad) presented through goggles featuring tiny screens.

So when you read popular accounts of subjects who had their brains scanned while they were shopping, for example, remember that they were not walking around a Wal-Mart with headgear on. Far from it.

Via Pure Pedantry. See also here.

The best experiments, I think, would be precisely those where people could walk around Walmart with headsets on and we could measure some of their brain activity. Sadly, that's not such an easy thing to do. What we do therefore is strap them into scanners or else conduct simulated experiments in the laboratory. While these are very valuable, the point is -- we cannot read too much into them.

Is there social science research on the habits of movie-renters?

I've just started exploring the Netflix data -- and want to use it to look into the habits of folks who rent movies regularly. There's a lot of work on collaborative filtering algorithms and other "recommendation" systems that goes on in Computer Science departments.

But does anyone know of any academic work that was done documenting the behavior patterns exhibited by movie-renters? Questions such as:

• Is the renting behavior exhibited by families different from that of individuals?
• How does liking or disliking a movie affect the next movie that person or family rents?
• What role does the social network play? What about advertisements?

Anything that anyone knows will be much appreciated.

In praise of skimming: A response to Nicholas Carr's "Is Google making us stupid?"

Nicholas Carr's essay in the Atlantic -- provocatively titled "Is Google making us stupid?"-- has, not surprisingly, aroused a lot passionate reactions.

Channeling Maryanne Wolf's Proust and the Squid, Carr makes the reasonable assertion that the way we read shapes the way we think. And increasingly, we do most of our reading on the internet: skipping, skimming and scanning our way through articles and blogposts. He contrasts this style of reading with the kind of deep contemplation that is necessary for reading, say, War and Peace. If reading on the internet discourages "deep" reading, Carr worries, does that mean we are all starting to think differently -- becoming more "transient" and less reflective? Is deep thinking is on the way out? Is Google making us stupid? (I know, I know, this sounds like one of those voiceovers from Sex and the City).

This is an interesting and important question. But Carr's article, at least in the way it is framed, manages to conflate a number of related issues. It's fundamental flaw is the implicit assumption that there is only one (correct) way to read and it involves "strolling through long stretches of prose". It doesn't matter whether we're reading fiction or non-fiction or whether we're reading for work or in our leisure time. The result is an essay whose underlying anxiety manages to resonate but whose point isn't at all clear. I actually agree with Carr. I certainly think that the internet is changing something in the way discourse producers work (and by this term I mean academics, journalists, writers etc who read and write for a living). Notice that this is a much narrower claim -- how they think is another, more nebulous issue and I am not particularly sure that that's going to change.

Consider the disconnect in Carr's first anecdote.

Bruce Friedman, who blogs regularly about the use of computers in medicine, also has described how the Internet has altered his mental habits. “I now have almost totally lost the ability to read and absorb a longish article on the web or in print,” he wrote earlier this year. A pathologist who has long been on the faculty of the University of Michigan Medical School, Friedman elaborated on his comment in a telephone conversation with me. His thinking, he said, has taken on a “staccato” quality, reflecting the way he quickly scans short passages of text from many sources online. “I can’t read War and Peace anymore,” he admitted. “I’ve lost the ability to do that. Even a blog post of more than three or four paragraphs is too much to absorb. I skim it.”

Notice that there are two arguably unrelated things here:

• (a) Friedman's inability to "read and absorb a longish article on the web or in print": This presumably has to do with his work as a blogger.
• (b) Friedman's inability to read War and Peace in his leisure time.

Carr clearly seems to be implying that (a) and (b) are related, and that both are caused by the amount of time that Friedman, a blogger, spends on the internet for work.

I suspect that reading War and Peace after a hard day's work is ... difficult, even for those whose work doesn't involve reading tons of things on the internet. Indeed, as Friedman admits here (scroll down for his comment) -- he reads a lot of novels (mainly mysteries and thrillers) for recreation in his leisure time. We can therefore safely eliminate Friedman's high volume of internet reading as the cause of (b).

(a) is much more interesting but Carr has next to nothing to say about it. Now presumably, as a discourse-producer, Friedman has to look at different books, articles, blog-posts, papers and then synthesize/summarize them in blog-posts of his own. But notice that when Friedman says that he cannot read a blog-post of more than 4 paragraphs, he means that he does not read it word for word, instead he skims it. In other words, when something slightly longish presents itself to him, Friedman slips, almost by default, into "skim mode".

Carr clearly despairs at this but my own reaction is: so what? Carr seems to think that all reading takes place in isolation -- that we read for the sake of reading. But while this may be true for our leisure reading, this is almost certainly not true when we read for work. The real issue here is that assuming Friedman skims more now than he did pre-internet, has it affected his output? Is his writing better or worse? Are his arguments sharper, deeper, shallower? Does he write more? Or less? These are all very interesting questions but it doesn't seem to me that Carr is really interested in these. (Of course, these questions might get you published in an academic journal but they certainly wouldn't be cover-story material for The Atlantic Monthly, no?)

So now let me restate Carr's point -- which he may or may not buy. His thesis is that the internet and Google will change the way that these producers of discourse go about their work, that they will skim more than they used to -- and more worrying for Carr, I think -- and they will not think as deeply because deep reading is deep thinking.

I am not convinced about the second part of this at all. First, because discourse producers don't just read aimlessly, they read for a purpose (an article, an essay, a review) and that purpose itself will require rigorous thinking. In the course of their project, they may go through a variety of reading material (blog-posts, journal papers, books). They may skim most of it, deep-read a few. But in the end, they have to synthesize their arguments and marshal facts and sources to back them up -- and to do all this well, deep thinking is essential.

However:

• The weak part of Carr's thesis is certainly true -- we do skim more. And perhaps reading on the internet may have reduced our "skimming threshold" -- like Friedman who starts skimming if an article is more than 4 paragraphs. But the reason for this is that the internet presents us with ten times the reading material we would otherwise have had.

In my own case, to write this blog-post, I read Carr's original piece in the Atlantic the "traditional" way, word for word. I skimmed through most of the reactions the article generated -- concentrating on the few that seemed most relevant. (Clay Shirky's is the most penetrating, if a bit short. Jaron Lanier's is bizarrely short and off-topic.) I knew the New Yorker had reviewed Proust and the Squid but after looking it up on Google, I ended up skimming four reviews: times.co.uk, Bookforum, The New Yorker and the Guardian. Google also led me Bruce Friedman's comment about his reading habits. And it took me to the study by University College, London, that Carr quotes from.

This is perhaps the key problem that discourse producers face -- what to skim and what to read deeply. And the internet has exacerbated it -- because now everything is available at our fingertips. To be successful today, it not only becomes necessary to skim but it becomes essential to skim well.

• We may end up writing in shorter bites -- almost certainly a good thing in my opinion.

As Kevin Drum points out, most non-fiction articles are grossly padded. Knowing that there will be a lot of articles users will want to browse, will writers try to be more concise in their arguments? More brief, to the point? I am sure this will not be a completely unwelcome development.

Conclusion? The internet certainly encourages skimming -- but skimming isn't easy! It could certainly mean less deep-reading in the long run but deep thinking will almost certainly remain. Finally the very fact that there is so much to skim will compensate to some extent the fact that we have less time to read something word by word.

The Kindle

James Fallows has a good post on the deficiencies of the Kindle, and by extension, most online reading applications:

3) And about the process of reading:

Spent six or seven hours of the flight reading on the Kindle. Perfectly pleasant and legible. Only one inconvenience relative to " real" books -- harder to flip ahead or back several pages at a time. (You scroll page by page, or else go to the table of contents.) And a kind of mental-picture adjustment: it's easier to insert bookmarks or placeholders, or seach for a specific word in the text; harder to have a remembered visual image of a certain passage as it fits on a certain place on a page. Not good for books where pictures, illustrations, maps, production quality matter a lot. Very, very good for reading Word .DOC files or .PDFs that I would otherwise have to read on the computer.

In the same vein:

One added observation, however, would be that the Kindle actually suffers from several ridiculous flaws. James refers to the inability to "flip" multiple pages at a time. It also doesn't let you cross-reference Kindle "locations" with brick-and-mortar page numbers. And you can only highlight whole lines at a time rather than starting with specific words. There are various other things like that. They're annoying. But at the same time, these are problems that I'm sure have solutions. When the basic technology of the Kindle Reader and Kindle Store are married to a design team (either at Amazon or at a competing firm like Apple) that's somewhat better at thinking this stuff through then I think you'll have a product a lot of people want to buy.