88
RICE UNIVERSITY STUDIES
4. Much more detailed studies of the time series associated
with library operations must be made. Analytical methods that
will permit the objective determination of the points of transi-
tion from periods of one type of growth to another must be de-
veloped as well as the interrelation of statistical information
pertaining to different types of growth statistics.
5. It seems clear that there is a connection between the growth
of the library archives and the growth of various estimators of
the state of civilization. It ought to be determined which is cause
and which is effect. More precisely, the role of the preservation
and transmission of information in the development of civiliza-
tion certainly ought to be investigated not only as a subject of
abstract intellectual interest but also to provide a working tool
for those responsible for allocation of national resources. Fur-
thermore, as the details of this relationship become clarified,
library management will be in a better position to improve the
means of accessing the information archive.
NOTES
1. Gelb (Ref. 4) writes: “If the alphabet is defined as a system of signs
expressing single sounds of speech, then the first alphabet which can justifi-
ably be so called is the Greek alphabet.”
2. The data given in Table II appears to agree with that used by De Solla
Price for his figure on page 115 of Reference 2, but his description of the
growth there given seems to be in error.
3. The datum of 1967 does not lie on the fitting line. Since the most recent
statistic in an economic time series is usually revised, this number has been
ignored in fitting the exponential.
4. The Kennedy confrontation with the steel industry concerning its pricing
policies occurred after the sharp transitional drop in BOP steel output growth
rate in 1960-61. The proposed price increases may have been a reaction to this
transition. Had it been known at the time that the transition was of Type-III
but not of Type-II and that it would last for one year in its depressive
phase, the industry might not have reacted with its proposed non-transitional
price increase. This example illustrates the importance of studying the
causes of transitions and learning how to distinguish the various types as they
occur.
This problem also illustrates the importance of further statistical study of
the procedures to be used in interpreting growth data. It is possible to “fit”
the BOP steel output data with a single straight line that effectively “hides”
the transition period of 1960-61. The procedures for testing the improvement
introduced by the notion of a break in the growth function are fairly obvious.
Isolation of the period involved, given that a break occurred, is also straight-
forward. However, the problem of finding break points when there is no
a priori information as to how many are present does not appear to have
been studied at any length.