|Reinventing discovery : the new era of networked science
by Michael A Nielsen
Publisher: Princeton, N.J. : Princeton University Press, 2012 .
In 1999 Microsoft organized a chess game between Garry Kasparov, who was then world chess champion, and ‘the world’. In a dedicated forum, self-selected players would propose and vote moves on behalf of the ‘world team’. Kasparov prevailed after 62 moves in a match where, as he declared, he expended ‘more energy than ever before’. The move n. 10, a ‘great contribution to chess’ according to Kasparov, was devised by Irina Krush. Albeit she was – and is – a talented chess master, she could not rival with Kasparov. Yet Krush had studied intensely move n.10 before and when the game serendipitously landed in move 9 in an apt configuration of pieces, she was just the right person at the right moment – and put Kasparov in dire straits. Krush’s appearance was indeed serendipitous – after all she might have decided not to take part to the forum – but the world team was carefully organized so as to make such serendipitous encounters between the right people – the ‘expert’ – and the right task not unlikely. This is what Michael Nielsen calls ‘organized serendipity’, a form of cooperation based on allocating expertise where is most needed by networking a great number of people with a broad array of capacities. Organized serendipity could – according to Nielsen – change the face of how research is done. This is however not the only enhancement of a networked science. The ingredients of discovery are human creativity and existing knowledge, and networks are making the storage and sharing of existing knowledge very effective. Open access repository of pre-print papers, or the sequence of the human genome in the public domain, are key examples of these opportunities. Nielsen envisages on these two bases a new mode of discovery that relies on information ‘commons’ and an efficient system of allocation of tasks to experts. Information and creativity are already there, networks will only make it easier to mobilize them and exploit the ‘cognitive surplus’ that otherwise would be wasted in dusty libraries, or leisure. ‘Online tools offer a fresh opportunity to improve the way discoveries are made, an opportunity on a scale not seen since the early days of modern science’.
Nielsen brings a number of examples to sustain his claims, including the Polymath project – which helps mathematician to address hard problems by collecting suggestions from the internet; practices of data sharing among scientists, among which those that have led to the establishment of the hap-map repository of common genetic variants; and citizen science projects, from Galaxy zoo to Foldit, respectively a crowd-powered galaxy taxonomical project, and a computer game that help researchers to predict protein 3D structure from their amino acid sequence. He also notices that although there are powerful barriers that may hinder the development of a ‘networked science’, including the current structure of incentives for scientists, things are rapidly changing with the spread of open-access publishing and even science blogging.
Remarkably for a book that squarely targets a general audience, Nielsen presents some details of the organizational model for research he has in mind, a model that is heavily indebted with Benkler’s theories of peer production (ref.1) although with a marketist spin that is altogether absent in the original. Nielsen is seemingly fascinated by Hayek’s description of how markets recruit information that would otherwise be lost (ref 2) and struggles with the notion of ‘design of attention’ to devise a system which could compare to market-based price signals in the allocation of experts to tasks that ‘maximise their comparative advantage’. However, the only convincing example he comes up with is InnoCentive, which is a recruitment system that relies on a embryonal pricing system – with prospective ‘employers’ offering rewards for problem solving, rewards which signals their importance. Nielsens’ recommendations are equally inspired to market models, especially where he recommends that tasks within the discovery process ought to be modularized in order to lower entrance barriers, and mechanisms to ensure approximation to perfect information.
There is one particular issue that is left unscrutinized, an issue on which at the end depends to what extent ‘discovery’ will be revolutionized by networks – as arguably other sectors of cultural production might be. Benkler identified a series of preconditions for networked production to emerge. These include the presence of a large ‘information common’ that everybody can access and diffuse infrastructures for making and disseminating one’s products. When the cost of sharing information was high, the production and diffusion of knowledge was monopolized by few agencies with enough capital to buy rotary presses and broadcasting antennas, namely states or big firms. To the extent Benkler’s conditions are met, we will be able to organize some chunks of theoretical science in an entire new manner, powered by the collaboration of experts and lay people outside restricted academic circles. Yet large bits of discovery still depend on extremely expensive hardware. Few companies and state-run institutions can afford DNA sequencers or the organization of clinical trials – let alone hardron colliders. Nielsen does a great deal to show that discovery may require big numbers rather than lonely geniuses. However he still overemphasizes the typical contribution of geniuses, creativity – and forget technologies and painstaking routine lab work. Networked science will certainly influence discovery, but it remains to be seen which fields will be most affected and how these novelties will impact on the rest of the research system – which may mutate less that Nielsen imagines and hopes.
(1) Yochai Benkler, Coase’s penguin, or Linux and the nature of the firm, Yale Law Journal 112(2002), 369–446
(2) Frederich A. Hayek. The use of knowledge in society. American Economic Review XXXV(1945), No. 4, 519-30