What Is an Emerging Technology?

1 Introduction

• The literature lacks consensus on what makes a technology emergent, with definitions emphasizing different features and analyst perspectives shaping judgments of emergence.Technologies grouped under broad labels may have distinct socio-technical features.Source: §1 · p. 2
• Measurement approaches are eclectic and often lack strong links to the concepts measured, producing divergent analyses even when similar methods are used.Changing database categories also make the phenomena examined by scientometric methods less clear.Source: §1 · pp. 2–3
• 12.5% (all disciplines) and 23.8% (social sciences) were the average yearly growth rates of emerging-technology publications from 1980 to 2013, versus 4.9% for all SCOPUS publications.Figure 1 also tracks news articles mentioning emerging technologies.Source: Figure 1 · p. 3
• The paper defines emerging technologies through five attributes: radical novelty, relatively fast growth, coherence, prominent impact, and uncertainty and ambiguity.Coherence persists over time; the technology may exert considerable socio-economic impact, but its most prominent impact lies in an uncertain future.Source: §1 · p. 4
• The operationalisation framework reviews indicators and trend analysis, citation analysis, co-word analysis, overlay mapping, and hybrid approaches, adding approaches from other fields where scientometrics is insufficient.The framework is intended to connect the proposed attributes with measurement options.Source: §1 · p. 4

2 The concept of emergence

• Emergence is a process of becoming rather than a static property, with dictionary definitions differing on whether existence alone or a degree of prominence is required.Endpoints are variously described as visible, evident, important, or prominent.Source: §2 · p. 5
• In complex systems theory, emergence can mean novel properties arising from interacting parts that are not reducible to those parts or predictable from them.Goldstein defines it as novel, coherent structures, patterns, and properties arising through self-organisation; de Haan distinguishes ontological properties from epistemological mechanisms.Source: §2 · pp. 5–6
• The authors argue that ‘emerging technology’ usually uses the standard sense of emergence, focusing on novelty, uncertainty, early identification, and prominence rather than system origins and causal interactions.Some technologies may be complex systems, while others are designed engineering systems that exhibit complicatedness without adaptation or self-organisation.Source: §2 · p. 6

• Radical novelty is contextual: an existing technology can be novel when applied in a new domain, while incremental improvement alone is not the intended meaning.The authors also exclude science-basedness as a defining requirement.Source: §3 · pp. 8, 10, 13
• Relatively fast growth is judged against technologies in the same domain, while coherence entails a connected community and a distinct identity sustained over time.Growth can be tracked across actors, funding, knowledge outputs, prototypes, products, and services.Source: §3 · p. 10
• Prominent impact may range from reshaping one domain, such as clinical practice for a disease, to affecting multiple domains or the wider socio-economic system.The impact should be contextualised within the domain or domains where the technology emerges.Source: §3 · p. 11
• Uncertainty concerns future impacts and their likelihoods, while ambiguity reflects fluid or contradictory applications and incomplete knowledge of possible outcomes; the two can coexist.The authors describe them as a continuum rather than mutually exclusive conditions.Source: §3 · pp. 11–12
• The attributes co-evolve: early novelty can precede growth and coherence, while later growth and impact may stabilize or decline as uncertainty falls and a community becomes established.The authors treat emergence as the phase of dramatic attribute change and note that observers may disagree about its boundaries.Source: §3 · pp. 11, 14
• The authors link the attributes to measurement, using scientometric techniques as the main approach and other approaches to address gaps.They frame operationalisation as necessary to classify technologies as emerging or not.Source: §3 · p. 15

4 A framework for the operationalisation of emergence

• A review narrowed 155 search results to 55 relevant studies, grouped into five methodological approaches that show definitions of emergence vary even within technique groups.The framework addresses contemporary and retrospective cases and notes that expert validation remains important; future-oriented technology analysis is excluded from the review, though relevant to prospective decision-making.Source: §4 · pp. 15–16

4.1 Radical novelty

• Radical novelty can be assessed contemporaneously through news and expert-facing publications, and retrospectively through citation, co-word, and overlay analyses of changing knowledge structures.These approaches have limits: contemporary materials may be difficult for non-experts to interpret, citation and co-word analyses require longitudinal data, and overlay maps rely largely on visual interpretation.Source: §4.1 · pp. 20–23

4.2 Relatively fast growth

• Retrospective fast growth is measured through document counts, fitted growth curves, bursts, diffusion, and expanding clusters, but contemporary publication and patent counts cannot yet reveal it.Counts require criteria to distinguish emergence from other rising trends, while diffusion classifications depend on selected thresholds and observation periods; funding, big data, and altmetrics are potential early signals.Source: §4.2 · pp. 23–25

4.3 Coherence

• Coherence signals an emerging technology’s acquired identity and persistence, measurable through narrowing terminology, shared acronyms and institutions, persistent or dense clusters, and increasingly connected co-authorship networks.Retrospective methods also assess entropy, cluster persistence, and convergence among conference sessions; patent property-function networks use small, dense sub-networks as a coherence signal.Source: §4.3 · pp. 25–27

4.4 Prominent impact

• For prominent impact, scientometrics is limited in contemporary cases, so qualitative and mixed STS approaches use contemporaneous expectations and visions to examine how technologies may reshape domains and socio-economic systems.Expectations can stimulate and coordinate action but also produce hype and contested claims about impact or coherence; expectation mapping can be paired with patent co-classification and qualitative analysis when data permit.Source: §4.4 · pp. 27–28
• Retrospective scientometrics can assess impact through highly cited documents, authorship and co-authorship patterns, and citation- or term-based cognitive-network dynamics (e.g. Hicks et al., 1986; Melin and Persson, 1996).Source: §4.4 · p. 28

4.5 Uncertainty and ambiguity

• Contemporary uncertainty and ambiguity can be assessed through news, editorials, reviews, and perspectives, while STS mapping of expectations can help assess uncertainty and identify multiple visions of the future.Source: §4.5 · p. 29
• Retrospective scientometric work on uncertainty remains sparse and mostly measures reduced uncertainty in scientific communication via category redundancy or entropy across titles, references, and time, not uncertainty about technologies’ impacts or uses.The assessment of uncertainty and ambiguity is identified as an important area for future research.Source: §4.5 · p. 29

• The five attributes used to characterise technological emergence are radical novelty, relatively fast growth, coherence, prominent impact, and uncertainty and ambiguity.Source: §5 · p. 30
• Scientometric methods’ suitability for operationalising emergence depends on time, the attribute, and the data used.Source: §5 · p. 30
• Scientometric techniques are better suited to retrospective than contemporary analysis because publications and patents take time to appear and be indexed.Growth is particularly difficult to measure contemporaneously; future-citation methods are more affected than co-word analysis and bibliographic coupling.Source: §5 · p. 30
• Scientometrics has little use for operationalising uncertainty and ambiguity, and reviewed methods do not explicitly assess the societal aspect of prominent impact.Source: §5 · p. 30
• Studies focused on publications and patents are time-sensitive and provide limited perspectives on the multifaceted phenomenon of emerging technologies.Source: §5 · p. 30
• Method choices can make detected emergence an artefact; time-window length can miss either slow-forming patterns or technologies emerging only briefly.Short windows may fail to capture a critical mass of documents, while longer windows may miss technologies that exhibit emerging features for a shorter period.Source: §5 · pp. 30–31
• Qualitative STS approaches can complement scientometrics by interpreting expectations and societal impact, while scientometrics offers statistical inference and greater generalisability.The paper notes that few studies combine the two traditions.Source: §5 · p. 31
• Future research could use full-text, funding, and real-time altmetric data, but these sources have limits in coverage, information, and comparability.Full-text analysis may reveal competing envisioned applications; funding growth may indicate fast growth, while acknowledgements identify funders but not funding amounts. The paper also calls for research on emergence’s origins, counterfactual cases, and endpoint.Source: §5 · pp. 32–33

• The literature lacks a widely accepted definition and conceptually grounded operationalisation framework, with differing views carrying implications for policy and governance.Source: §6 · p. 34
• Emerging technologies are defined by five attributes: radical novelty, relatively fast growth, coherence, prominent impact, and uncertainty and ambiguity.The definition describes a relatively fast-growing, radically novel technology with coherence persisting over time and potential socio-economic impact whose most prominent effects lie in an uncertain and ambiguous future.Source: §6 · p. 34
• The framework links attributes to measures: document counts over time assess growth, new clusters indicate novelty, and entropy or new categories assess coherence.The paper also discusses the connectivity of new clusters as relevant to assessing novelty; examples include Porter and Detampel, 1995; Kajikawa and Takeda, 2008; Small et al., 2014; and Cozzens et al., 2010.Source: §6 · pp. 34–35
• Scientometric assessment of uncertainty and ambiguity and prominent impact remains largely unexplored, and publication and patent data may not suit contemporary cases.The paper also identifies the risk that apparent emergence is an artefact of selected models.Source: §6 · p. 35
• Qualitative STS analysis of news, reviews, and policy documents can illuminate uncertainty, ambiguity, and prominent impact, especially in contemporary cases.STS interpretation may also reduce the likelihood of false positives or missed patterns in scientometric results.Source: §6 · p. 35
• Future research should examine emergence’s origins, compare technologies that do and do not emerge, and investigate when emergence ends.The paper also identifies full-text, funding, altmetric, and broader big-data sources as opportunities to improve assessment of attributes that current scientometrics addresses only to a limited extent.Source: §6 · p. 35
• The paper presents conceptual clarification as a precondition for coherent operationalisation, future research, and more informed policy-making and governance.Source: §6 · p. 36

• Table A1 lists studies of the concept of emergence in complex systems theory chronologically.Source: Table A1 · p. 44