Monday, September 4, 2017

Much is written, but little is understood

I can't find this on the internet but it is quite interesting. From The New Scientist, 9 May, 1992.

Elsewhere, I comment a lot on the bubble effect - groups who erect barriers between themselves and the rest of the world. The bubbles can originate in physical barriers but are usually social in nature - affiliation by class or income or religion or neighborhood or profession, by accent, etc. This research suggests there is an unconscious incentive structure in place to create communication barriers between researchers and the general public and scientists from other fields.
Much is written, but little is understood
by William Brown

Scientific journals are becoming harder and harder to read according to the first scientific analysis of the language they use. Increasingly "incomprehensible" articles are alienating the public and cutting scientists off from important research in neighbouring fields.

Donald Hayes of Cornell University, New York, has found that all the leading scientific journals have grown steadily harder to understand since the turn of the century. "Science has become inaccessible," he says.

In Hayes's ratings, difficult articles receive a positive score, easy articles a negative score. Newspapers on average score 0, while comics score —26. New Scientist scores an easy-to-read 4.

In the 1930s, Nature and Science, the most prestigious journals for publishing research, had scores close to 0. They were as easy to read as newspapers and research papers were accessible both to other scientists and to the general public. But by 1950, Science scored 5 and Nature scored nearly 20. By 1980, Science scored 15 and Nature over 25. By 1990, Science was scoring 28 and Nature 31.

Hayes finds the same trend in all the leading specialised journals, from biology to astronomy. "Even researchers working in related fields cannot read these articles any more," he says. "The articles are just too hard."

Hayes rates articles by examining the frequency with which authors use common, and not so common, words. Each of the 10,000 most common words in English is ranked according to its frequency in every-day use. For example, the word "the" is top of the list. To estimate the difficulty of an article, Hayes counts the number of times the article uses each word.

The simple language of comics includes few rare words and many common words. Difficult scientific articles include a lot of obscure words and jargon. "Research papers leave out most of the words ranked between 100 and 1000," says Hayes.

The cause of the difficulty is the increasing degree of specialisation in science. "Specialisation has brought great rewards," Hayes says in his report in last week's Nature. "But the broad consequences are that ideas flow less freely across and within the sciences, and the public's access to (and maybe trust in) science is diminished."

But Hayes also speculates that journals may have deliberately made their articles harder. "It seems that editorial policy may have had something to do with it," he says.

John Maddox, editor of Nature, acknowledges that research papers in the journal have become harder to read over the years. But this reflects the way scientists write, not an editorial decision, he says.

"Scientists have ceased to care about communication," says Maddox. "They are more concerned with publication, just to keep their jobs. If you are dismayed you should read the papers when they come in. They are often changed entirely out of recognition."

Mike Green, a leading expert in string theory, one of the most abstract theories in particle physics, believes it is impossible to communicate advances in his field in ordinary language. "It is not possible to explain theories of quantum gravity at a level which a biologist could criticise," he says.
In Obfuscating ideologically unacceptable research findings, I observed that there were multiple motives for obfuscation. Principally:
A) complexity, B) unacceptability, or c) mediocrity
I would posit, that in at least some of the fields, candidate c is the dominant factor.

Increasing communication complexity is a quite interesting finding, regardless of the root cause.


UPDATE: The readability of scientific texts is decreasing over time by Pontus Plavén-Sigray et al. Same finding but from a much larger data set and brings it up to the present. The trend continues.

No comments:

Post a Comment