Thursday 6 December 2007

Gossip test & Boredom principle

Crick’s gossip test and Watson’s boredom principle: A pseudo-mathematical analysis of effort in scientific research.

Bruce G. Charlton. Editorial. Medical Hypotheses. Volume 70, Issue 1, 2008, Pages 1-3

Summary

Crick and Watson gave complementary advice to the aspiring scientist based on the insight that to do your best work you need to make your greatest possible effort. Crick made the positive suggestion to work on the subject which most deeply interests you, the thing about which you spontaneously gossip – Crick termed this ‘the gossip test’. Watson made the negative suggestion of avoiding topics and activities that bore you – which I have termed ‘the boredom principle’. This is good advice because science is tough and the easy things have already been done. Solving the harder problems that remain requires a lot of effort. But in modern biomedical science individual effort does not necessarily correlate with career success as measured by salary, status, job security, etc. This is because Crick and Watson are talking about revolutionary science – using Thomas Kuhn’s distinction between paradigm-shifting ‘revolutionary’ science and incremental ‘normal’ science. There are two main problems with pursuing a career in revolutionary science. The first is that revolutionary science is intrinsically riskier than normal science, the second that even revolutionary success in a scientific backwater may be less career-enhancing than mundane work in a trendy field. So, if you pick your scientific problem using the gossip test and the boredom principle, you might also be committing career suicide. This may explain why so few people follow Crick and Watson’s advice. The best hope for future biomedical science is that it will evolve towards a greater convergence between individual effort and career success.

***

The gossip test

“It came to me that I was not really telling [people] about science. I was gossiping about it. This insight was a revelation to me. I had discovered the gossip test – what you are really interested in is what you gossip about.”

Francis Crick. What mad pursuit: a personal view of scientific discovery, 1988 [1].

The boredom principle

“...Never do anything that bores you. My experience in science is that someone is always telling you to do things that leave you flat. Bad idea. I’m not good enough to do well something I dislike. In fact, I find it hard enough to do well something that I like.”

James Watson. Succeeding in science: some rules of thumb, 1993 [2].

The two most famous co-discoverers of the structure of DNA, Francis Crick and James Watson, gave complementary advice to the scientist who wishes to do the best work of which they are capable. The crux is that to do your best work you need to make your greatest possible effort.

Crick made the positive suggestion to work on the subject which most deeply interests you, the thing about which you spontaneously gossip – Crick termed this ‘the gossip test’; Watson made the negative suggestion of avoiding topics and activities that bore you – which I have termed ‘the boredom principle’.

This is good advice because science is tough. The easy things have already been done. Solving the harder problems that remain requires a lot of effort: either a lot of hours of investigative work, or a lot of hours of thinking – or sometimes both.

Effort must be sustained. Unless effort is fuelled by interest it will not last long enough to solve the problem, if effort is pushing against a counter-current of boredom it will be weakened. The story of the discovery of DNA’s structure is one of a massively concentrated joint effort over a relatively few years, but even over this unusually short time span there were many serious setbacks – enough to deter anyone who was too-bored or insufficiently-fascinated by the problem [3].

The effect of gossip test and boredom principle on Individual Effort can be expressed pseudo-mathematically as follows:

Let G equal the amount of time spent on gossiping about the subject that really interests you, and CP represent the time spent gossiping about the ‘current project’ on which you are supposed to be working. The percentage of maximum effort of which you are capable then equals the ratio of CP divided by G.

Individual Effort=CP/G


This is the gossip test of individual effort
However, all science has a certain percentage of boring aspects. The boredom principle could be framed to state that the percentage time spent on activities which are boring in your current project (CP) must be subtracted from the CP effort. This percentage is the ‘boredom quotient’ – BQ. Therefore:

Individual Effort(percentage of your maximum possible effort)=(CP/G)–BQ.


This equation places a bogus, but superficially-impressive, quantification onto Crick and Watson’s insight.
It must acknowledged, however, that individual effort in modern biomedical science does not closely correlate with career success as measured by salary, status, job security, etc. This is because Crick and Watson are talking about revolutionary science – using Thomas Kuhn’s distinction between paradigm-shifting ‘revolutionary’ science and incremental ‘normal’ science [4]. Modern biomedical research is overwhelmingly ‘normal’ science – indeed, in organization and structure it resembles industrial R&D (research and development).

There are two main problems with pursuing a career in revolutionary science [5]. The first is that revolutionary science is intrinsically riskier than normal science because you are less likely to succeed. The second is that even success in triggering a paradigm-shifting revolution in a scientific backwater may be less career-enhancing (generating less status, salary and job security) than mundane work in a trendy (=well-funded) field.

So, if you pick your scientific problem using the gossip test and the boredom principle, you should indeed give yourself the best chance of making a personal contribution to the achievement of a major breakthrough. But the statistical probability of actually achieving a breakthrough remains small – the bigger the problem, the tougher to solve. And you might also be committing career suicide, by working in a low status, poorly funded, scientific backwater.

These aspects can be included in a new equation for measuring the probability of ‘career success – CS’. This modifies the input of Individual Effort by introducing two extra phony-variables: 1. percentage ‘probability of solution’ of the problem (PoS) and 2. ‘professional status’ of the field (PS).

Revolutionary science has a much lower PoS than normal science – leading to a lower probability of CS. And the gossip test and boredom principle will often direct individuals to work in fields where the PS is sub-optimal – also leading to a reduced CS.

So we arrive at the equation:

Percentage likelihood of career success CS=(CP/G)–BQ×PoS×PS


where CP is the time spent gossiping about current project; G is the time spent gossiping about favourite topic; BQ is percentage of boring activities in CP; PoS is probability of solution of the problem; and PS is the percentage professional status of that branch of science as reflected in the proportionate funding, journal impact factors, number of jobs compared with the trendiest area.
This contrived equation yields the insight that the course of action leading to the greatest level of career success may differ substantially from the course of action leading to the highest probability of achieving a breakthrough in revolutionary science.

All of which may explain why so few people follow Crick and Watson’s advice. Implicitly the majority of scientists are not seeking their best chance of contributing to major breakthroughs in revolutionary science; but instead are seeking to optimize their career success by pursuing normal science in trendy fields. It also explains why so few people put 100% of the maximum possible level of effort into their current project – because they are working in areas which contradict the gossip test and boredom principle.

The best hope for future biomedical science is that it will evolve a convergence between Individual Effort and Career Success. Nothing can be done to alter the greater riskiness of revolutionary science compared to the predictability of incremental R&D. But maybe it is not unreasonable to hope that revolutionary science will increase its professional status [5].

References

[1] F. Crick, What mad pursuit: a personal view of scientific discovery, Penguin, London (1988).

[2] J. Watson, Succeeding in science: some rules of thumb, Science 261 (1993), p. 1812. View Record in Scopus | Cited By in Scopus (2)

[3] H.F. Judson, The eighth day of creation: makers of the revolution in biology, Penguin, London (1979).

[4] T.S. Kuhn, The structure of scientific revolutions, Chicago University Press, Chicago (1970).

[5] B.G. Charlton, Why medical research needs a new specialty of ‘pure medical science’, Clin Med 6 (2006), pp. 163–165.

Component-oriented scientific writing

How can the English-language scientific literature be made more accessible to non-native speakers? Journals should allow greater use of referenced direct quotations in ‘component-oriented’ scientific writing

Bruce G. Charlton

Editorial. Medical Hypotheses. Volume 69, Issue 6, 2007, Pages 1163-1164 - doi:10.1016/j.mehy.2007.07.007

Summary

In scientific writing, although clarity and precision of language are vital to effective communication, it seems undeniable that content is more important than form. Potentially valuable knowledge should not be excluded from the scientific literature merely because the researchers lack advanced language skills. Given that global scientific literature is overwhelmingly in the English-language, this presents a problem for non-native speakers. My proposal is that scientists should be permitted to construct papers using a substantial number of direct quotations from the already-published scientific literature. Quotations would need to be explicitly referenced so that the original author and publication should be given full credit for creating such a useful and valid description. At the extreme, this might result in a paper consisting mainly of a ‘mosaic’ of quotations from the already existing scientific literature, which are linked and extended by relatively few sentences comprising new data or ideas. This model bears some conceptual relationship to the recent trend in computing science for component-based or component-oriented software engineering – in which new programs are constructed by reusing programme components, which may be available in libraries. A new functionality is constructed by linking-together many pre-existing chunks of software. I suggest that journal editors should, in their instructions to authors, explicitly allow this ‘component-oriented’ method of constructing scientific articles; and carefully describe how it can be accomplished in such a way that proper referencing is enforced, and full credit is allocated to the authors of the reused linguistic components.

***

In scientific writing, although clarity and precision of language are vital to effective communication, it seems undeniable that content is more important than form. Potentially valuable knowledge should not be excluded from the scientific literature merely because the researchers lack advanced language skills.

Given that global scientific literature is overwhelmingly in the English language, this presents a problem for non-native speakers, especially for those whose language differs markedly from English in terms of its basic grammatical structure. This has become a particularly acute problem with the exponential expansion of Chinese science with an annual doubling of publications from this source [1]. Because, although many non-English speaking scientists are able to acquire sufficient competence to understand the English scientific literature; it is much more difficult – sometimes impossible – for them to learn how to write English with sufficient clarity and precision for effective scientific communication.

The traditional practice has been for scientists either to employ a translator – which is expensive and may not be possible – or to rely on line-by-line sub-editing services to be provided by the scientific journals – which is also expensive and is not possible for all journals. Furthermore, detailed sub-editing is very time-consuming if done well, and if done badly may end by significantly distorting the intended expression of ideas. And anyway, unless the submitted paper reaches a certain standard of linguistic comprehensibility, it will be rejected and will never even reach the stage of being sub-edited.

My proposal is that scientists should be permitted to construct papers using a substantial number of direct quotations from the already-published scientific literature – whenever the author judges that these quotations are a precise and clear exposition of what s/he would like to say – if only they had the linguistic competence. Naturally, such quotations would need to be explicitly referenced so that the original author and publication should be given full credit for creating such a useful and valid description.

At the extreme, this might result in a paper consisting mainly of a ‘mosaic’ of quotations from the already existing scientific literature, which are linked and extended by relatively few sentences comprising new data or ideas.

Such a result could not be regarded as an ideal for scientific writing; on the other hand it may be the best attainable result (from the perspectives of clarity and precision) which is possible within the existing constraints of the real world – and that is surely sufficient. The result should, at any rate, be better than insisting that scientists of poor linguistic competence be compelled to re-phrase and re-combine concepts which have already been well-expressed elsewhere in the scientific literature.

This model bears some conceptual relationship to the recent trend in computing science for component-based or component-oriented software engineering – in which new programs are constructed by reusing programme components, which may be available in libraries [2]. A new functionality is constructed by linking-together many pre-existing chunks of software. Implicit is the notion that it makes sense to reuse functional units when they have proved effective in the past – the same could apply to the principle of reusing functional units of English language, which have previously proved effective in expressing standard scientific concepts.

I suggest that journal editors should, in their instructions to authors, explicitly allow this ‘component-oriented’ method of constructing scientific articles; and carefully describe how it can be accomplished in such a way that proper referencing is enforced, and full credit is allocated to the authors of the reused linguistic components.

Acknowledgement

Thanks are due to Peter Andras for the example of component-oriented software engineering.

References
[1] Zhou Ping and Loet Leydesdorff, The emergence of China as a leading nation in science, Research Policy 35 (1) (2006), pp. 83–104.

[2] Wikipedia. Software componentry. /wiki/Software_componentry; 2007 [Accessed 04.07.07].