Next month, I am scheduled to give a short speech (three to five minutes in length) at the annual induction ceremony of the American Academy of Arts and Sciences in Boston. This is a bit different from the usual scientific talks that I am used to giving; there are no projectors, blackboards, or other visual aids available, and the audience of Academy members is split evenly between the humanities and the sciences (as well as people in industry and politics), so this will be an interesting new experience for me. (The last time I gave a speech was in 1985.)
My chosen topic is on the future impact of internet-based technologies on academia (somewhat similar in theme to my recent talk on this topic). I have a draft text below the fold, though it is currently too long and my actual speech is likely to be a significantly abridged version of the one below [Update, Oct 12: The abridged speech is now at the bottom of the post.] In the spirit of the theme of the talk, I would of course welcome any comments and suggestions.
For comparison, the talks from last year’s ceremony, by Jim Simons, Peter Kim, Susan Athey, Earl Lewis, and Indra Nooyi, can be found here. Jim’s chosen topic, incidentally, was what mathematics is, and why mathematicians do it.
[Update, Nov 3: Video of the various talks by myself and the other speakers (Emmylou Harris, James Earl Jones, Elizabeth Nabel, Ronald Marc George, and Edward Villela) is now available on the Academy web site here.]
— Introduction —
If I had to name the most significant technological development in recent decades, I would have to say it would be the internet. By this, I mean not just the physical architecture of the internet per se, which was already available to academics and government agencies since the 1960s, but also all the innovative technologies that flourished once the internet matured, from tools as humble as the email mailing list to such unreasonably effective services as modern search engines or Wikipedia.
As the internet has become more integrated into the mainstream of modern life, it has disrupted and revolutionised one sphere of human activity after another. We read in the news about how online media is thriving as “old” media stumbles; how online medical information is transforming patient-doctor relationships; how blogs, tweets, and online videos are tipping the balance in closely fought elections; and so forth.
But to most of us in academia, there is a temptation to view these changes with a certain detachment: sure, established for-profit companies may well face competition (as they ought to) from lower-cost internet-based rivals, and it is only reasonable in a democracy that politics should be influenced by popular debate, both offline and online, but we, by contrast, should be secure in our ivory towers from any internet revolution, with our tenure, our unique expertise, and our time-tested academic traditions.
Even when new technologies do hit close to home – by threatening the profit model of the academic journal system, say, or by greatly facilitating the ability for students to cheat on their homework (and also for professors to detect such cheating!) – we can still rationalise away these developments as requiring only superficial changes to adapt to – switching from physical journals to online journals, perhaps, or placing more safeguards on our homework formats. We still perform our “core” academic activities – teaching, advising, research – much as we have for over a century: classroom by classroom, student by student, and paper by paper. We may do more of these things online now rather than offline, but it is still the academic who is at the center of things, not the internet. After all, it is not as if our classes can be replaced by a Wikipedia entry, or our research by a search engine query, right? Right?
Well, yes and no. It’s true that even the most advanced online resources available today are not nearly “smart” or sophisticated enough to render our academic services obsolete; not yet, at least. Unlike many other industries, academia does not currently face any real threat from a cheap internet-based competitor.
But I believe a “hybrid” form of academic activity is beginning to emerge – one in which internet-savvy academics and their institutions harness the full power of online tools to initiate and organise large research collaborations, and to disseminate and share their results at far more rapid and effective rates than were previously possible. In my discipline – mathematics – this type of net-centric activity is still in its infancy, but it shows signs of potentially being substantially more efficient (and perhaps more importantly, open, cumulative and responsive) than traditional collaboration and dissemination, and is likely to become increasingly mainstream in the years ahead. It may not totally revolutionise the way we work, the ambition of what we hope to achieve, and the academic culture we work in, but it is likely to transform them significantly.
— Teaching —
Consider teaching, for instance. Year after year, day after day, and in universities across the world, we stand in lecture halls and present the foundations of our subject to classrooms consisting of hundreds, or even just dozens, of students at a time. This keeps us engaged with our students, hones our skills, and makes us feel useful, but is it the most efficient way to do things?
There is a mathematical topic – Mobius transformations – which is taught routinely in complex analysis classes in a thousand mathematics departments across the world, to classes of perhaps thirty or fifty students in size; I have done so myself several times. On Youtube, there is a beautiful video explaining the geometric interpretation of these transformations which has been viewed one million, six hundred thousand times so far – more people than can be reached than by even ten thousand mathematics lecturers. It can be accessed by just about anyone on the internet through a simple web search on the topic (it is in the top three hits currently on all major search engines).
Now, clearly, one cannot hope to replicate the entire classroom experience as a sequence of Youtube videos – the quality of interactivity, depth of material, and availability of expert attention, in particular, is much poorer. Even more professional organised efforts, such as the online videotaped lectures offered by institutions such as MIT, are an imperfect substitute for physically being present at these lectures. But the sheer numbers of people one can reach by the internet shows the potential of tapping this medium to teach in the future.
Already, hundreds of academics (including myself) use a blog to post their course notes and encourage online discussion (in all directions) between the teacher and students in the classroom, as well as visitors from around the world; I have had classes with perhaps thirty local students but up to a hundred other participants from a variety of backgrounds following (and commenting!) using the blog. There is a much higher quantity and level of questions asked, and the material in my notes is much improved, because of this; and I have learned more about the subject than if I had taught it in a traditional way, both from preparing the blog material, and from obtaining feedback from students and participating colleagues.
Even after the physical class ends, the online class goes on; I have often had people wanting to learn a subject stumble onto one of my online lecture notes on my blog from a year ago through a search engine, and continue the discussion afresh. Within a few years, there may well be valuable online content like this for virtually every commonly taught academic topic, just one search query away from anyone with internet access.
The technological level of online interactivity is certain to increase in the future; one can well imagine it becoming routine in classes to (for instance) field questions by text message from students overseas who are watching the lecture in real time through video, with the discussion continuing online long after the class has ended. Not all experiments in online teaching will achieve their intended objectives, but it only takes one clear success to provide a model that can then be rapidly emulated by institutions and lecturers worldwide.
In my view, the traditional classroom lecture will still play an indispensable role in the future, but in a rather different format than it is today, with its effects being vastly amplified and prolonged through its integration with the internet.
— Collaboration —
Another major area where profound changes are happening is that of collaboration in research.
It was only four decades ago that the primary mode of communication among academics in distant institutions was by physical mail. This was inconveniently slow, and it discouraged collaboration with anyone who was not in the same physical location. With modern communication tools such as email, the situation today is vastly different; it is completely routine now in mathematics to collaborate over long distances, with months of online communication punctuated by only a few (but crucial!) days of physical contact each year. Perhaps as a consequence, there has been a huge increase in the proportion of papers in mathematics that are jointly authored, rather than singly authored. As a related phenomenon, an increasing fraction of papers are also interdisciplinary rather than specialised to a single subfield.
Very recently, software tools have become available to allow easier collaboration by large numbers of authors from across the world. Unlike the sciences, pure mathematics in academia has never really had the large laboratories in which armies of graduate students, postdocs, and senior researchers work on a single goal; but the technology is just becoming available for such large-scale projects to be possible.
This year, for instance, by ad hoc usage of existing tools such as blogs and wikis, the first “polymath” projects were launched – massively collaborative mathematical research projects, completely open for any interested mathematician to drop in, make some observations on the problem at hand, and discuss them with the other participants.
The very first such project solved a significant problem in combinatorics after almost six weeks of effort, with almost a thousand small but non-trivial contributions from dozens of participants. It was a novel way to do mathematics, but also a novel way to locate the collaborators with the right expertise and interest to solve the problem, perhaps serving as a model to begin collaborations through online networking rather than physical networking.
And there were other unexpected benefits too; the projects have retained a fully available online record of all the discussion, including false starts, dead ends, and incremental progress, that took place while the problem was not yet solved, giving a much richer, more dynamic, and more accurate picture of how mathematical research really takes place than the cut-and-dried presentations one sees in finished products such as papers and textbooks.
By taking research online, it comes to life; one participant compared his anticipation to seeing the latest developments on a polymath project to the suspense one might feel while watching a TV or movie drama. Veteran researchers are familiar with these tensions, frustrations, and joys, but it used to be quite difficult to convey these experiences to the graduate students entering the field; perhaps these open internet projects, with their “show, don’t tell” nature, may succeed in doing so in the future.
— Academic culture —
As we adopt new technology, our culture of doing things subtly changes. In mathematics, for instance, research used to be a secretive activity; one would often not discuss what one was working on before it was ready for submission to a journal, and would only give out preprints to a select few colleagues before the publication process was complete (which takes months or even years). With the rise of preprint servers and search engines, it is nowadays quite customary to put a preprint online as soon as it is submission-ready (or sometimes even sooner!); experience has shown that doing so greatly increases the visibility, impact, and influence of one’s work, and (perhaps counterintuitively) discourages excessively competitive behaviour and even plagiarism, as the timestamps given by preprint servers can help defuse arguments over precedence.
Indeed, in many parts of mathematics there is now a social expectation that one’s work should be readily available online, and journals have largely abandoned attempts to enforce a (counterproductive) monopoly on the dissemination of their authors’ work. As a result, research developments propagate at a significantly faster speed than in previous decades.
In the future, I can imagine further cultural shifts of this type. Currently, the actual problem-solving process in mathematical research is usually obscured from view until the problem has been solved and a polished, publication-quality draft is available; with the rise of open collaborative projects such as polymath, this culture may begin to change in the future. (For instance, I circulated a draft of this talk on my blog weeks in advance, both to obtain valuable feedback and to encourage me to continue working on the text. A few years ago, I might only have shown a draft to one or two trusted friends, with perhaps a single round of revisions.)
Similarly, the advent of mathematical blogs and other semi-formal outlets for discussion is reinforcing an existing trend in mathematics in which the intuition and motivation behind a mathematical topic is emphasised as much as the definitions, theorems, and proofs; some of the more technical and specialised subfields of mathematics may well encounter increasing societal pressure in the future from their peers to make their work more accessible and transparent to wider audiences.
In teaching mathematics, the current model is that of a nearly one-way street; the lecturer does almost all of the talking. Apart from a few questions from the more bold students, one only receives feedback days or weeks after the class has ended, from the assignments, evaluations and exams the students turn in.
With improvements in technology, there may be a greater expectation in the future for such classes to be significantly more interactive, both during the “actual” class, as well as the online discussions before and afterwards, and with near-instant feedback becoming the norm.
Such changes will certainly encounter resistance; consider for instance the ongoing debate on whether to allow laptops in classrooms. Many such initiatives will not be fully successful; we still have a very partial understanding of what makes one online experiment flourish and another one fail. Nevertheless, I doubt that we will keep the status quo indefinitely in the presence of such technological and social changes.
— Conclusion —
One can draw an analogy between pre-internet academia and pre-industrial manufacturing. Before the industrial revolution, manufacturing was the province of individual craftsmen or of secretive guilds, working painstakingly on each individual piece of work, with each master passing down their carefully hoarded insights and tricks to just a handful of disciples. It is not hard to find parallels to each of these phenomena in academia.
But after the industrial revolution, specialisation and mass production became the paradigm in manufacturing; less intimate, surely, but also vastly more efficient and reliable. One might bemoan the loss of creativity and individuality that each craftsman exhibited, but eventually, as the industrial revolution matured into the modern era, the outlets for creativity became dispersed to a wider group of people. Thanks to division of labour, design, invention, entrepreneurship, manufacturing, marketing, training, or management could now be performed by whoever was best qualified to do each, rather than by the same individual; and the best practices in each of these areas could be adopted widely, rather than being confined to their originator and a select number of followers.
Academia has not experienced change on the scale of the industrial revolution since the invention of the printing press. With the advent of the internet – the modern day analogue of the printing press, among other things – could it be revolutionised once again?
— Abridged version of speech —
It’s a great honour, both to be inducted to the Academy and to address you all today. I must confess that while I have given over a hundred scientific talks, this is only my second speech; and the first one was when I was nine. So I please bear with me; I’ll try not to sound like a nine-year-old.
I would like to talk about the impact of the internet, and all the unreasonably effective services it has spawned, from modern search engines to Wikipedia.
We know that the internet has revolutionised area after area: entertainment, journalism, politics will never be the same again. But those of us in academia like to feel protected in our ivory towers from the internet revolution, with our tenure, our expertise, and our academic traditions. After all, our classes can’t be replaced by a Wikipedia entry, and our research can’t be replaced by a search engine – not yet, anyway.
Nevertheless, I believe major change is already underway.
Consider teaching, for instance. There is a mathematical topic – Mobius transformations – which is taught in a thousand mathematics departments across the world, to perhaps thirty or fifty students at a time. I’ve done so myself many times.
But if you do a web search for Mobius transformations, you’ll find a beautiful video on Youtube explaining this concept clearly, which has been viewed one million, six hundred thousand times – more people than can be reached by ten thousand mathematics classes.
On a smaller scale, hundreds of academics (including myself) have actively pushed their classes onto the internet, using such tools as blogs. I have had classes with perhaps thirty local students but up to a hundred online participants. Even after the physical class ends, the online class goes on, with new visitors stumbling onto the class via a search engine and continuing the conversation.
These tools can have unexpected uses; for instance, I posted a draft of this talk online a few weeks ago, and got a tremendous amount of valuable feedback in return.
Or consider research. This year, for instance, by ad hoc usage of existing tools such as blogs and wikis, the first “polymath” projects were launched – massively collaborative mathematical research projects, completely open for any interested mathematician to drop in.
The very first such project solved a significant problem in combinatorics after almost six weeks of effort, with almost a thousand small but non-trivial contributions from dozens of participants. It was a novel, transparent, and lively way to initiate and then do mathematics. One participant even compared his anticipation to seeing the latest developments on a polymath project to the suspense one might feel while watching a TV or movie drama. (You had to be there, I guess.)
Academia has not experienced massive change – on the scale of the industrial revolution – since the invention of the printing press. With the advent of the internet – the modern day analogue of the printing press, among other things – could it be revolutionised once again?