Some patterns of technological development
Using the sliding window approach outlined in Section 5.4, 13 temporally specific networks were constructed. The earliest one includes patents from 1987-1994. The latest has patents from 1999-2006. The sliding window was moved one year from one temporally specific network to create the next temporally specific network. Using the structural property of broad technological areas meant that each network has six vertices. The lines between them represent the citations between technological categories. The newly constructed variable, vertex size, was saved as a vertex property. Four of these technologically defined networks are shown in Figure 5.7.
However, before visually examining details of these short networks, it is useful to consider the growth in the number of patents by technological categories over time. These are shown in Figure 5.6 for 1987-1999 (summed over the eight-year time intervals defined in Section 5.4.1). Computers and communications (C&C) had the second lowest number of patents in 1987. However, it had the steepest rise in the number of its patents, finishing with the highest number of patents in 1999. Electrical and electronics (E&E) started at the third lowest level in 1987, had the second largest increase in patents and finished with the second
Figure 5.6 Growth in the number of patents by categories from 1987 to 1999.
highest number of patents. Together, these two categories dominated both the growth of patents issued during this period and the absolute levels of granted patents towards the end of it. Undoubtably, this was driven by the development of the computer industry, its rapid emergence, and its derivatives.
While drugs and medical (D&M) started at the lowest volume of patents and remained there, it had the third steepest rise through 1995 before leveling off. The sharp early rise was likely due, in part, to an increasing use of medications to handle a variety of problems, especially psychological problems. Another reason could be the aging population of the USA; disproportionate amounts of medical costs in the USA are incurred in treating the aged in efforts to improve the quality and length of their lives. See, for example, Berk and Monheit (1992, 2001); Conwell and Cohen (2005); Yu and Ezzati-Rice (2005). These two goals need not be compatible. The treatments include greater uses of medications and surgery. There was also the rise of sports medicine. Both mechanical and others had the highest levels of patents in 1987 but grew only slightly before finishing third and fourth in total patent volume in 1999. Chemical (Chem), one of the drivers of technological change following WWII, barely grew through 1995 before leveling off and then declining slightly.
Figure 5.7 shows four networks with citation volumes between technological areas, each defined by the eight-year width of the sliding window. The first covers 1987-1994, the second 1991-1998, the third 1995-2002, and the final one 1999-2006. The changes in vertex sizes across the four panels in Figure 5.7 are fully consistent with the trends shown in Figure 5.6. The number of granted patents increased throughout this period. This accelerated for C&C, E&E, and D&M (except at the end). More interesting here are changes in the flows between technological areas, as shown by the thickness of the arcs. They represent the total number of citations from one technological category to another. Of course, citation volumes are highest within technological areas, sometimes by several orders of magnitude. These could have been included as loops. However, given our focus on citation flows between technological areas, loops were removed from Figure 5.7.
The most noticeable features of Figure 5.7 are the reciprocal arcs between C&C and E&E. They were the largest flows in 1987-1994 and roughly equal in volume. Both grow in every time slice. These changes were particularly dramatic in the last two periods. After 1987-1994, the citation volume from C&C to E&E exceeded the reverse flow. For 1999-2006 they are, by far the dominant flows as they were for 1995-2002, albeit to a lesser extent. Although most of the citation flows between technological areas increase across the four periods, further evidence of the 'patent citation dominance' of C&C and E&E comes from considering the flows between them and the other technological areas.
While mechanical remained stable with regard to patent citation volumes, its reciprocal flows with both C&C and E&E increased across all periods. In 1999-2006, these were the second largest reciprocated flows. Indeed, the flows between these three technological areas grew to be the largest in 1999-2006. The citation volumes between mechanical and others were the second largest in 1987-1994 and grew thereafter. It appears that patents in mechanical drew more heavily on patents in others in the last two periods than the reverse flow. The reciprocal flows between chemical and others grew. Also of interest is the growth in the reciprocal flows between chemical and D&M through the first three periods before diminishing. In terms of patents, this suggests that the initial growth for D&M shown in Figure 5.6 may be driven more by the increased use of medications.
Figure 5.7 Shrunken networks in the four year time windows for citation flows.
Loops were removed. Given the large volumes and vertex measures, line values were multiplied by 0.0002 and vertex values by 0.002