Convergence

The preceding chapter examined how patent activity distributes across CPCCooperative Patent Classification — a hierarchical system jointly managed by the USPTO and EPO that categorizes patents by technology area (e.g., H = Electricity, G = Physics). technology fields and how those fields have grown at different rates. Yet the boundaries between technology fields are not fixed. As digital technology has become pervasive, inventions increasingly span multiple CPC sections, and once-separate domains have become deeply intertwined.

This chapter examines cross-domain convergence from two complementary perspectives: the aggregate trend toward multi-section patents, which captures the overall rate of convergence, and the specific pairs of CPC sections that account for that convergence, which reveals where technology boundaries are most permeable. Together, these measures demonstrate that the patent system is no longer a collection of independent silos but an increasingly interconnected network of technological activity.

The convergence matrix reveals that cross-field integration is not uniformly distributed. The G-H (Physics-Electricity) pair has dominated convergence from the late 1990s onward, reflecting the deep integration of computing, electronics, and physics that characterizes the digital era. In the earliest period (1976–1995), B-C (Operations-Chemistry) and A-C (Human Necessities-Chemistry) pairs led convergence, but the pattern shifted substantially as digital technology became pervasive.

The share of patents spanning multiple CPCCooperative Patent Classification — a hierarchical system jointly managed by the USPTO and EPO that categorizes patents by technology area (e.g., H = Electricity, G = Physics). sections has increased over time, indicating growing technological convergence. Contemporary inventions increasingly draw on knowledge from multiple domains, a characteristic of the digital era in which software, electronics, and traditional engineering fields intersect.

The aggregate rise in multi-section patenting can be decomposed into two distinct sources using a shift-share framework. Within-firm convergence captures the extent to which individual organizations have broadened their technological scope over time, filing patents that span more CPC sections than they did previously. Between-firm convergence reflects compositional shifts in the population of patentees — for instance, the growing share of firms that are inherently multi-domain (such as large technology conglomerates) relative to single-domain specialists. Understanding whether convergence is driven primarily by firms individually expanding their technological breadth or by structural changes in the composition of patenting organizations has important implications for innovation policy and competitive strategy.

Interdisciplinarity in patenting can be decomposed along two distinct dimensions:scope (how many technology domains a single patent spans) andconvergence (how frequently patents combine specific pairs of domains). These dimensions capture different aspects of cross-domain innovation and have distinct implications for the complexity and novelty of inventions.

A natural concern with the convergence findings is whether they reflect genuine technological change or artifacts of the CPC classification system. To assess robustness, this section compares multi-section patent rates under both the CPCCooperative Patent Classification — a hierarchical system jointly managed by the USPTO and EPO that categorizes patents by technology area (e.g., H = Electricity, G = Physics). and the older IPCInternational Patent Classification — a hierarchical classification system managed by WIPO that categorizes patents by technology area. The IPC forms the basis for many national systems and is closely related to the CPC. classification systems. If the convergence trend is driven by CPC-specific reclassification or taxonomy changes, it should not appear under IPC.

The first four chapters established how the patent system grew, which technology fields accompanied that expansion, and how those fields increasingly converge across once-separate domains. The analysis now turns to the language of patents themselves. The shift from chemistry to digital technology is visible not only in formal classification codes but in the words inventors use to describe their work. The Language of Innovation applies unsupervised text analysis to five decades of patent abstracts to uncover the latent thematic structure of US patenting.