Wissenschaftswachstum in wichtigen naturwissenschaftlichen Disziplinen vom 17. bis zum 21. Jahrhundert

Abstract

Wissenschaftswachstum in wichtigen naturwissenschaftlichenDisziplinen vom 17. bis zum 21. Jahrhundert Summary: The Growth of Science within Important Scientific Disciplines from 17th to 21st century. -In this paper the growth of knowledge is investigated for the disciplines of Astrophysics / Astronomy, of Physics and of Chemistry in the last centuries. In this context the main emphasis is devoted to the discussion of the growth of literature. Besides, this is also illustrated by the growth of indicators like for example important books, new discoveries in science, the chemical elements, the known chemical compounds and the number of known inorganic chemical structures. Compared to numerous earlier studies a systematic evaluation of all presently available data sources is carried out. The data are fitted by different mathematical growth models (linear, quadratic, exponential, hyperbolic). These are discussed and compared. As far as the exponential growth model is applied the doubling periods obtained have values between 11 and 32 years for the cumulative total of publications, between 24 and 98 years for that of important books, between 79 and 163 years for that of new discoveries and between 10 and 14 years for that of chemical substances.From an order of magnitude point of view this is in agreement with the results of other authors. It is however also demonstrated that the exponential growth model which is normally used is not necessarily the best to describe the growth of the data. Other models are often better suitable to explain the growth. It is shown that linear growth of the yearly publications describes the data better than exponential growth in certain epochs. In this latter case it is demonstrated that the parameter (linear slope) for the rate of increase has a constant values in certain epochs, but rises in steps from epoch to epoch by a factor of 3 to 40. After the second world war the linear growth model for the yearly publications or the quadratic growth model for the cumulative total of publications respectively delivers the best results. This is therefore used for forecasting the future.

Mathematics 1868–2008: a bibliometric analysis

“…This means that the doubling time of the growth of the literature in mathematics in the 19th century is about 19 years. By the way, this is in very good agreement with the corresponding growth rate in physics in the 19th century where the doubling time is 20 years, but faster than in astronomy/astrophysics where the corresponding doubling time was 33 years and slower than in chemistry with a doubling time of 14 years (Behrens and Lankenau 2006). As mentioned above, the ZMATH Online Database covers the literature from 1868 until today.…”

“…In this context, the model of linear growth of the annual number of publications or of the quadratic growth of the cumulative publications, respectively, provides also an excellent description of the data after the year 1960. By the way, this is also valid in the case of physics, chemistry, astronomy/astrophysics and crystallography (Behrens and Lankenau 2006, Behrens and Genz 2008, Behrens and Luksch 2006. For forecasting the future by applying extrapolations, this latter model should be preferred.…”

“…(May 1963). On the other hand, it should be mentioned that the growth rates of the literature in physics (doubling time 11 years), in astronomy/astrophysics (doubling time 18 years), chemistry (doubling time 14 years and crystallography (doubling time 11 years) were faster in the second half of the 20th century (Behrens and Lankenau 2006, Behrens and Genz 2008, Behrens and Luksch 2006.…”

A bibliometric study in crystallography

“…The usual method is by a fit of an exponential function in the time range 1935-2004 (shown here) with R 2 = 0.989 and a doubling time T = 10.6 years. This doubling time is in agreement with earlier results (Behrens, 1996) and approximately in agreement with the doubling time of the growth of the whole chemical literature, which has a value of 14 years (Behrens & Lankenau, 2006). The second method uses a fit of a thirddegree polynomial with a coefficient of determination R 2 = 0.9997, also in the time range 1935-2004 (shown in Fig.…”

“…This assumption was originally suggested by de Solla Price (1963). However, as explicitly discussed by Behrens & Lankenau (2006), there are many cases in which it is not justified. Bradford's (1934) law describes the distribution of the literature for a given subject in periodicals (journals).…”