Friday, May 28, 2010

How we read (articles and magazines) now

In this post, I want to compare the two modes of reading: how we read before the internet and how we read now. I will be limiting the analysis only to news articles and magazines since I believe, pace Nicholas Carr, that it isn't still clear how the internet has changed book-reading habits. But today, the internet IS the platform for delivering news and I believe it has changed our reading habits substantially. (It's also caused a big disruption in the publishing industry.)

Just to be clear, the analysis in this post is based on a sample size of 1: me! But I believe the salient points will hold true for many others.

How we read Pre-internet

Back in the days when there was no internet, we had two newspapers delivered to our home in the morning: a newspaper in English (we alternated between The Times of India or the Indian Express) and one in Marathi (Maharashtra Times or Loksatta). On Mondays we also took in the Economic Times and on weekends, my father would usually buy some more from the news-stand (The Sunday Observer, The Asian Age).

My father would usually read every newspaper from the first page to the last. My mother would usually only read the Marathi newspaper. My sister and I would usually at least skim through the English newspaper but concentrate on a few sections (usually comics, editorials, and sports).

Even in India, where newspapers are generally smaller, (i.e. less number of pages compared to, say, the tome that is the daily New York Times), a newspaper would try to cover everything it thought needed to be covered. So while the bulk of it would be devoted to national politics, there would be a smattering of world news, book reviews, entertainment and sports news, etc. No issue would be covered in too much depth (because of lack of space, which in turn corresponded to the high cost of newsprint) and the issues covered would be chosen keeping in mind the broad preference of a newspaper's readers.

So if you were interested European Union politics, you would be lucky to get an article every two weeks. But at the same time, the very fact that newspapers felt that they were the only way we got content, they did feel obligated to at least sample the whole spectrum of topics. E.g. even if the sports pages were dominated by cricket, soccer and tennis, at least once in a while, volleyball and football would be covered.

In the figure above (click to see full figure), I have chosen to represent how we read back then in terms of two orthogonal axes: the breadth of reading (X-axis) and the depth of reading (Y-axis). The breadth of reading corresponds to the topics that we could cover, the depth of reading corresponds to how much we could cover on a certain topic. Each line represents a topic; how long the line is corresponds to how deeply a topic is covered.

The characteristics of pre-internet reading were:
  • The newspaper/magazine decided what topics it would cover (and it tried to take the interests of the majority of its readers into account). This meant, to some extent, that the "long tail" of topics was not covered.
  • The limited space available for a newspaper meant that:
    • The depth that it could go into for any topic that it covered was limited.
    • However topics were sampled uniformly across the spectrum of topics. Which meant that even if you didn't care for national politics at all, the very act of glancing through the newspaper everyday forced you to have some idea of what was going on. Ditto for sports, or for entertainment.
Pre-internet reading thus tended to emphasize breadth over depth. This was its advantage as well as its disadvantage. Diligent readers were "well-rounded" but they came up against the hard limit of newsprint if they wanted to know more.

How we read now

Reading on the internet is different.

There's far far more material to read, and it's easy to get to, requiring nothing more than a click of the mouse.

There's also more control over the material meaning you can decide what you want to read and ignore the stuff that you don't care for. E.g. if you are interested in economic policy you can subscribe to the Business feed of the NYT and nothing else. If you're interested only in foreign policy and particularly in the U.S.-China relationship, you can use Yahoo Pipes and set up filters that will bring to you only those articles that mention the U.S. and China.

Finally, there's more depth. A newspaper, even an online newspaper, can only devote so much time and space to U.S.-China relations. But on the internet, this material can be augmented with the many blogs and wikis out there. There are blogs by anthropologists and economists, by foreign policy and international relations experts, and by political scientists, -- you name it! -- all of whom are interested in conveying their point of view to both lay and specialized audiences. In other words, you can choose to specialize in whatever you want and your specialization will be world-class.

In the figure above (click to see full figure), I've tried to represent the characteristics of online reading in terms of its breadth and depth. The characteristics of internet reading are
  • We get to decide what topics we want to read about.
  • There is unlimited "space;" it is possible to access any conceivable magazine or newspaper; in addition there are blogs, wikis and Twitter feeds that we can use as our filters to get to the interesting stuff that we care for.
    • One can go as deep into a topic as one wants.
    • But since our time is ultimately limited, emphasizing depth has the effect of sacrificing breadth. One could go as deep as one wants and read about international relations but the cost could well be no time to read any sports news. Whether this is a good or bad thing, I am not sure.
Reading on the internet tends to emphasize depth over breadth. This is its advantage as well as its disadvantage. It's possible to indulge your interests a lot, but it comes at the cost of being "well-rounded."

Wednesday, May 12, 2010

Ranking and Sorting: two forms of cognition

There are two ways of looking at cognition.
  1. There is cognitivism which looks at the mind as an information-processing system. In this view, it is thinking, i.e. the ability to play with representations of the world inside the mind, that is all-important.
  2. Then there is another view, a heterogeneous set of views that is sometimes called the situated action perspective. Some of the proponents of this view would be Phil Agre, Lucy Suchman, Edwin Hutchins, Rodney Brooks and Eleanor Rosch. In this view, it is action that is all-important. Abstract rule-based thinking is but one, and by no means all-important, resource for action. Equally important are our use of tools and artifacts, and our taken-for-granted bodily and social skills. The situated cognition view locates cognition in our interactions with the world rather than in thinking inside the head. Representations are key but they are by no means solely in the head nor do they mirror the world; rather they are resources for action and are embodied in the tools and artifacts that we use for our tasks.
If you know me (or read this blog), you probably know that I find the situated action perspective far more convincing,. I can't, by any means, give any kind of a logical "proof" for it. But I do think it provides a better description of how we do things in our day-to-day life. E.g. in a form I filled out recently, I was asked to rank a list of 16 items. From a cognitivist perspective, this is a "problem" that is "solved" using utility-maximization; what I hope to do in this post is to show the actual processes that I used to rank the alternatives and also to show that these processes are better described as my interactions with certain external representations of the task. These functioned as tools i.e. resources that helped me complete the task. I also hope to show that by transforming the problem from one representation to another, I could have done the task differently (and more easily).

The housing form is shown above and it asks me rank the 16 types of housing available. Each type, as you can see in the figure, is an ordered pair consisting of a residence hall, and a type of apartment. I was supposed to assign ranks from 1 to 16. I could choose not to rank an option (thereby signifying that I wasn't interested). And I could give multiple options the same rank, if I wished.

From the cognitivist perspective, the process by which I actually ended up ranking these options is not very important. Because at an abstract computational level, my actions can be described as a simple case of utility maximization. I have a certain set of preferences (an upper-limit for the rent I am willing to pay, an interest in living with room-mates rather than alone, etc.). Based on this, whatever I do can be described as calculating a value for each option that measures its suitability for me and then ranking the alternatives based on said values.

But consider what actually happened as I was trying to fill up this form.

It was almost impossible to just straightforwardly rank 16 items without messing up. Some of the problems that I ran into were:
  1. It was impossible to go through the alternatives one by one and assign them a rank. There were two possibilities.
    1. I could go through option 1, give it a rank, go to option 2, and so on. This would only work if I had worked out all the ranks in my head before-hand (impossible) or if I had worked them out on pen-and-paper (which I hadn't).
    2. I could scan all the alternatives, find the one that I liked the most, give it rank 1, then scan all the remaining alternatives, find the one I liked second-best, give it rank 2. This didn't work because (a) I quickly forgot what rank I was on, (b) I had to keep in mind all the others I had previously ranked (or get them by quickly scanning all alternatives) and finally (c) I just didn't know scanning the alternatives which one I preferred.
  2. The form did not give all the information that I thought relevant. E.g. there are 2-bedroom or 3-bedroom apartments in certain residence halls which did not have a living room; I wanted to rank these options lower. Then there were options that I just did not prefer i.e. I had a "black-list." I had to flip back-and-forth between this page and the pages describing the apartments in detail so that I would not accidentally rank something in my black-list.
In a sense, the problem was this: not only was I trying to rank all the alternatives, I was also evaluating the criteria I was using to do so at the same time. That is, I was creating the "rules" that I would use to prefer one alternative to another in the process of applying them to specific alternatives! Note however that the representation that I had been given to work with -- a list of alternatives along with the possibility of assigning a number i.e. the rank to each -- was not very conducive to what I was doing.

Ultimately I ended up following the strategies below:
  1. I first determined a rough order of preference in my mind. E.g. (1) efficiency apartments, (2) 2-bedrooms, (3) 3-bedrooms, (4) 4-bedrooms, and (5) 1-bedrooms. Of course, this wasn't the whole story since there were many exceptions (the black-list) to this preference. But as a rough guess, it was accurate.
  2. Then I went through the options one-by-one keeping the above list in mind.
    1. So I first went through all the efficiency apartments (there were 4), looked at their particulars (their price, whether they were on my black-list, etc.) and ranked them. If one was in my black-list, I chose not to give it a ranking.
    2. Then I went through the 2-bedroom apartments and did the same. I had to keep the last ranking I had assigned in mind. E.g. if I had ranked the efficiency apartments from 1 to 3 (not ranking the one in my black-list), then I had to remember to rank the 2-bedrooms starting with the rank 4. It was crucial to remember this. Because if I forgot, I had to scan the whole list again for the last ranking so far and this was not a fun task. If I mis-remembered, then I would mess up all subsequent rankings. I could have chosen to write this last ranking down to avoid this.
    3. And then I had to do the same for all the other types of apartments: 3-bedrooms, 4-bedrooms and 1-bedrooms.
    4. One of the reasons why I could do this was because there were no "anomalies" in my choices. E.g. there was no stray 3-bedroom that I preferred to a 2-bedroom. Had that been the case, I would have had to adopt a slightly different strategy. As I was ranking the 2-bedrooms, I would have had to look at the 3-bedrooms as well. And I would have had to look at the 3 and 4-bedrooms together. And so on.
  3. Once all this was done, I went back to all the alternatives I had not ranked (because they were on my black-list). I had the choice of not ranking them or giving them the last few ranks. Again it was crucial for me to remember the last rank that I had assigned. E.g. if the last ranking was 13, and I had 3 more left, I would assign them to ranks 14, 15 and 16.
Remember that this is a gloss on how I did the task. A narrative that I have constructed by looking back on what I did and trying to be conscious of what I was doing when I did it. It should, by no means, be considered a true account of what I did.

But that said, let's look at this account and see whether it can be reasonably described in terms of rule-following or utility-maximization. The answer is clearly, at some level, yes. But consider the following points:
  1. While I did follow a rule (a "heuristic"), at no point did I assign any value for each alternative. Nor did I ever consider all the alternatives together.
  2. Instead I used one of my selection criteria to break down the 16 alternatives into 5 groups in decreasing order of preference (efficiencies, 2, 3, 4 and 1-bedrooms). Then I ranked the items in each group. This was far easier to do since each group had only 4 or 5 members.
  3. A lot of what I did was, in effect, my way of interacting with the available representation of the problem (in this case, a list whose items could be assigned ranks).
Now consider how much more simple the problem would have been if I had an interface that allowed me to sort the table, rather than rank it, using mechanisms like drag-and-drop.

Or suppose I had made written each alternative on a piece of paper, and then sorted the pieces into a list with my most preferred alternative at the top and the least-preferred one at the bottom. Like so.

This makes the task so much simpler, no?

In effect, this is a transformation of the task: from a ranking task to a sorting task. Computationally, the "problem" remains the same, but this transformation makes it easier to solve. Here is why sorting is easier than ranking:
  1. Sorting allows us to focus on a subset of the alternatives without being distracted by the others . E.g. I can focus on whether I prefer an Edgerton 1-bedroom or a Tang 3-bedroom by moving the representations of these alternatives close together and moving all the rest out of my field of vision. This is impossible to do in a static list whose items need to be assigned ranks.
  2. Sorting converts the ranking problem into an ordering problem. E.g. once I have arranged alternatives A, B, C in that order and I decide that I prefer C to B, all I need to do is insert C between A and B. The order changes immediately to A, C, B. If I was assigning numbers to A, B and C this same step would involve me having to change the ranks of each one of them.
  3. Sorting eliminates any need for memory -- which is the hardest for the mind to do without external representations. E.g. one does not have to remember, say, the last rank that one assigned; instead the representation will itself tell us what the current state of the matter is.
What I write here has similarities both to the distributed cognition hypothesis and the extended mind hypothesis.

Selected References
  1. Edwin Hutchins, Cognition in the Wild, New edition. (The MIT Press, 1996).
  2. Abigail J. Sellen and Richard H. R. Harper, The Myth of the Paperless Office, 1st ed. (The MIT Press, 2001).
  3. David Chalmers and Andy Clark, “The Extended Mind,” Analysis 58 (1998): 10-23.

Friday, May 7, 2010

A reply to Sam Harris

[A friend of mine enthusiastically linked to this Sam Harris piece on Facebook. I found it dreadful and philosophically incoherent, even on its own terms. I thought I'll post what I wrote about it in my email to my friend. I've tried to keep the post contained and my point is to show only that Harris point breaks down even if one accepts all his assumptions about morality.]

I was not suggesting that science can give us an evolutionary or neurobiological account of what people do in the name of "morality." Nor was I merely saying that science can help us get what we want out of life. Both of these would have been quite banal claims to make (unless one happens to doubt the truth of evolution or the mind's dependency on the brain). Rather I was suggesting that science can, in principle, help us understand what we should do and should want -- and, perforce, what other people should do and want in order to live the best lives possible. My claim is that there are right and wrong answers to moral questions, just as there are right and wrong answers to questions of physics, and such answers may one day fall within reach of the maturing sciences of the mind. As the response to my TED talk indicates, it is taboo for a scientist to think such things, much less say them in public.

Sam Harris wants not just a descriptive scientific theory of morals but a prescriptive one, a Science of Morals that helps people take moral decisions, but not just any decisions. It enables them to choose the right alternative where "right" is defined in some objective way. In other words, he says he wants to discover (or invent, or at the very least believes in the possibility of) an algorithm that helps people take the correct decision.

Now what would this algorithm be like? Are moral decisions subjective or objective?

Here Harris brings in John Searle. Searle says that subjective and objective, both of these words, have two senses, epistemological and ontological. So according to him, a scientific theory of consciousness (which doesn't exist yet) has to be epistemologically objective (meaning that the theory uses "inputs" that are accessible to any interested person) but the content of consciousness is still ontologically subjective (meaning only I have access to my states of consciousness).

So a scientific theory of morals would be like a scientific theory of consciousness: it would be epistemologically objective, meaning any interested person can get to the same conclusion if they applied it. But is the decision produced by this moral theory an ontologically objective one? Or is it ontologically subjective like conscious states?

Harris says that the scientific theory of morals should take into account states of consciousness and should promote those indicate the "well-being" of a person. Apparently advances in neuroscience and a better scientific understanding of consciousness will help us understand "well-being" neurobiologically i.e. scientifically i.e. objectively.

Therefore, we can have not just a descriptive scientific theory of morality but a prescriptive theory too. It will work like this. It will allow a person to make decisions that maximize his or her well-being. And since future developments in neuroscience will allow us to define "well-being" objectively, in terms of a person's conscious states, our prescriptive theory of morals will be objective also.

Well, not so fast.

If well-being is a conscious state, is it ontologically subjective or objective? If I know my Searle correctly, Searle thinks that ALL states of consciousness are ontologically subjective. So an algorithm that helps me make moral decisions depends on well-being, and well-being is ontologically subjective - how is a prescriptive theory of morals going to be ontologically objective?

Harris will probably say that this is wrong, a theory of consciousness will result in a definition of well-being that is ontologically objective. He gives this analogy in support:
As I said in my talk, the concept of "well-being," like the concept of "health," is truly open for revision and discovery.
So "well-being" is like health. I can prove someone is ill, even if he continues to deny it, by shoving a thermometer into his arm-pit and pointing out to him that his body temperature is higher than normal. A scientific theory of consciousness will allow us to produce thermometer-like instruments that can measure a person's well-being objectively, and thereby open the way to an ontologically objective prescriptive scientific theory of morals.

Whew. But this is all too abstract. Let's take specific cases and see how well this theory holds up.

Suppose I have a moral dilemma: should I take a job which lets me help people (which I want to do)? Or should I get a job which lets me do research on cognition (something that's not really going to help people, but which I would love to do)? Clearly I cannot do both. How will Harris' proposed theory help me? Presumably, it will calculate my well-being were I to choose any one of these alternatives and then tell me the one that maximizes my well-being.

But consider, dear reader, whatever I decide, I will miss something, yes? And that something will affect my state of well-being, no? If I decide to do research in cognition, what makes this a better moral choice? But there's more. The algorithm has to calculate my future well-being if it can help me make a choice. How does it do that? If I do choose a career as a researcher in cognition on the algorithm's recommendation and I hate it (because I don't get along with my colleagues), isn't the algorithm wrong?

But you will probably say that I am missing the point.

Harris doesn't want a theory that helps us make these day-to-day personal decisions.

He wants a theory that can yield results like these: (1) Not wearing a veil is objectively better than wearing a veil because it promotes a better state of well-being. (2) Liberal secular democracies are objectively better than theocracies because the "well-being" promoted by the former is greater than that promoted by the latter. (3) The "well-being" of religious people is objectively less than that of atheists or non-religious people. You get the drift.

I wish with all my heart that a theory that proves these things objectively will come into existence. But this reveals all the weaknesses in Harris' piece.

First of all, why should a prescriptive theory of morals apply only to these situations and not to my day-to-day personal decisions? After all if well-being can be defined in an ontologically objective way, then it should also apply to ALL decisions that affect my well-being, no?

But second, how is this prescriptive theory going to account for a woman who chooses to wear a burqua outside? And insists that she feels better with it. Harris would presumably drag her, kicking and screaming, to his well-being center, wire up her brain and demonstrate triumphantly to her that she really isn't well, despite her assertions to the contrary.

Harris would be far better off trying to define his "well-being" concept in terms of concrete things. E.g. by using statistics that demonstrate the better health, mobility, longer life, etc. of women who choose not to wear a veil to those who do. But that's not "objective" enough for him (it isn't, but I would be fine with that, it is far more coherent). People like him are always looking for "deep" reasons to demonstrate conclusions that are dear to them. And because Harris is a neurobiologist, he thinks some future scientific explanation of consciousness (derived, of course, from neuroscience) will help him with that. Well, best of luck with that!

[By the way, I see many parallels between this and the Marxist notion of "false consciousness."]

Monday, May 3, 2010

Interactive computing systems and the relationship between embodiment and teachability

In this post, I am going to talk about the notion of "embodiment" in human-computer interaction. In particular, I will explain how my understanding of the term has changed and how experiences with teaching my Dad how to use certain programs (Picasa, the Indian Railways online website, etc.) revealed to me another facet of the term: the relationship between the embodiment and teachability of an interactive system. If one wants to teach someone how to perform a certain task on an embodied interactive system, then written instructions (without pictures or visual aids) are almost always insufficient. In other words, one measure of the embodiment of an interactive system can be found by looking at the efficacy of written instructions in helping a novice perform a certain task.

Paul Dourish's "Where the Action is" is possibly one of my favorite books. The book's thesis is that as computers have developed, our interactions with them have changed in nature and have progressively become more "embodied." Dourish divides the history of computer systems into 4 successive time-periods based on the mode of interaction:
  • Electrical: To program a computer, one had to rewire its hardware.
  • Symbolic: Abstraction was introduced and separated hardware from software. Abstraction meant that coding got progressively simpler: from machine language to assembly language to high-level Fortran-like languages.
  • Textual: This refers to the development of command-line interfaces. These, for the first time, made interacting with the computer, seem like a "conversation."
  • Graphical: Finally, there was the development of the graphical user interface (GUI) with its desktop metaphor and 2-dimensional arrangement of files and icons. The 2-dimensionality of the GUI meant that users were able to exploit "further areas of human ability as part of the interactive experience." This meant the use of faculties such as peripheral attention, pattern recognition and spatial reasoning.
Dourish sees each successive stage as involving more and more of the distinctively "human" capabilities i.e. those skills that are most used in our interactions in our everyday life with other human beings. Embodied interaction, as Dourish defines it, therefore means an interaction that involves more and more of distinctively human skills. These could be bodily skills (like pointing, gesturing, moving, pattern recognition) or social skills (our workplace habits, our everyday assumptions, etc.). In other words, embodied interaction is a move towards integrating more and more of our real-world practices (at home, at work, at play and so on) in our interactions with computing systems.

There is another aspect to embodied interaction that has slowly come to my attention as I have been trying to teach my Dad to use computer software (email, Picassa, certain websites, etc.). It involves what I call its teachability.

Let me give some background. My father worked during a time when computers were not such a ubiquitous part of the everyday work environment. In particular, he worked at a time when there were special people who did computer work -- and therefore not everybody needed to use the computer. More so, the computer was used for what can be called high-tech scientific stuff; it wasn't used at all for the mundane things at work: sending emails/memos, for filling up your time-sheets, submitting vouchers for reimbursement, etc.

Consequently, my Dad never interacted with computers in any sustained way when he worked. But he was certainly aware of them. However this was back in the times when the command-line (Dourish's 3rd stage) was the primary mode of interacting with computers; MS-DOS ruled. When my Dad did start interacting with computers in a sustained way however (primarily to keep in touch with me here in the US), he was dealing with the GUI, a new and much more embodied stage of human-computer interaction.

Consequently he would ask me to write down instructions for him on how to do a certain action. I would agree but I would find the writing of instructions to be extraordinarily hard. Here's an example of what I mean. To copy a file from one directory to another in DOS is a simple command:

Now of course, even here, there are variations. You don't have to specify the directory if you want to copy something from or to the same directory you are in, etc. But the command itself is fairly simple and easy to understand.

Now consider doing the same thing in Windows and there turn out to be quite a few ways to copy a file from one folder to another (drag-and-drop, Edit menu, Context menu). I'm going to consider one of them here, to illustrate its complexity:
Go to the folder you want to copy from. Select the file and copy it (this can be done either by right-clicking the file and selecting "Copy" or by selecting the file and selecting "Copy" from the Edit menu). Then go to your target folder and paste the file there (again, this can be done in two ways, right-clicking in the target folder and selecting "Paste" or selecting "Paste" from the Edit menu).
Just the sheer complexity of what I wrote above makes it clear that in terms of giving written instructions, the command-line interface is far simpler than the GUI and involves less tacit assumptions (the phrases in red font, above).

Talking about the GUI involves an almost taken-for-granted use of metaphors. Consider, for example:

Go to the folder/Be in the folder/Be in some application: Go to a folder implies that the folder is a place. Go in a folder implies that the folder is like a house which you can go inside. I also ask him frequently to be inside some folder or to be inside some application.

Select the file: This is not really a metaphor although it comes up frequently in my discussions with my father. Many times, when I instruct him over the telephone to "select the file," he doesn't get what I mean or asks me whether I mean to left-click or right-click on the file icon. I also frequently ask him to "select the text" (e.g. to copy a hyper-link from Skype to Firefox to open a link that I sent him)

How does one teach people to use embodied interactive systems (GUIs), particularly when they have very little experience with computers, don't use the computer frequently and most important, are distant from us so we can only communicate with them through spoken and written instructions? Some points and observations:

Written instructions without pictures are almost completely useless.

Written instructions with pictures (screen-shots) can be useful.

However, it is best to give verbal instructions especially during the task itself*. That way, your instructions can be followed in real time and you will get immediate feedback about their intelligibility (i.e. whether or not they're understood) and efficacy (whether or not they worked).

Having a shared referenced visual object is even better. Meaning that if you are teaching someone how to use the Indian Railways Ticketing website, it is best to have it open in front of you and to carry out an instruction yourself so that you know exactly what the person you are instructing is looking at. This is easy to do for websites but much harder if you are instructing someone about how to use Picasa or how to copy files from a flash-drive to the hard-disk.

Which leads to the best possible and most embodied form of instructions. Screen-sharing along with a voice communication channel is the best way of all! This way, you know exactly what the other person sees and the instruction itself can be modified depending on how you see it being interpreted. E.g. if you ask someone to copy a file to a certain folder and this doesn't seem to get through, you can immediately supply them with instructions on how to copy files and take them through it step-by-step.

I guess the broader point that I am trying to build to is this: embodied interaction uses many bodily and social skills, and these are almost always tacit; it is hard to translate them into words. Verbal instructions (spoken and written) need to be backed up with visual aids like screen-shots and screen-sharing. Therefore it follows that one rough measure of the embodiment of an interactive system can be found by measuring the efficacy of written and spoken instructions, especially when the student and the teacher are distant, and the student is a novice at using computers.

*Whether your instructions will be remembered is another question.