A brief history of risk

Abstract

Despite increasing life expectancy and high levels of welfare, health care, and public safety in most post-industrial countries, the public discourse often revolves around perceived threats. Terrorism, global pandemics, and environmental catastrophes are just a few of the risks that dominate media coverage. Is this public discourse on risk disconnected from reality? To examine this issue, we analyzed the dynamics of the risk discourse in two natural language text corpora. Specifically, we tracked latent semantic patterns over a period of 150 years to address four questions: First, we examined how the frequency of the word risk has changed over historical time. Is the construct of risk playing an ever-increasing role in the public discourse, as the sociological notion of a 'risk society' suggests? Second, we investigated how the sentiments for the words co-occurring with risk have changed. Are the connotations of risk becoming increasingly ominous? Third, how has the meaning of risk changed relative to close associates such as danger and hazard? Is risk more subject to semantic change? Finally, we decompose the construct of risk into the specific topics with which it has been associated and track those topics over historical time. This brief history of the semantics of risk reveals new and surprising insights-a fourfold increase in frequency, increasingly negative sentiment, a semantic drift toward forecasting and prevention, and a shift away from war toward chronic disease-reflecting the conceptual evolution of risk in the archeological records of public discourse.

Keywords: Content analysis; Danger; Ngram Corpus; Public discourse; Risk; Topic model.

Fig. 1

Historical change in the frequency and sentiment for the word risk and its close semantic neighbors in the Google Books Ngram Corpus. (A) Frequency of risk, fear, danger, and hazard in the Google Books Ngram Corpus and frequency of risk in the Corpus of Historical American English (COHA). (B) Frequency of risk in five languages—English, Italian, Spanish, French, and German—in the Google Books Ngram Corpus. German is presented in a separate box because the frequency of risk is much lower in German than in the other languages. (C). Change in the sentiment for words co-occurring with risk, fear, danger, hazard, and death. Sentiment was adjusted to mean score of all words, such that valences > 1 indicate a more positive context than average. The word death is included to provide a sentiment benchmark, as its meaning and sentiment have remained stable over history.

Fig. 2

Semantic drift of risk, danger, fear, and hazard from 1800 to 2000 in the Google Books Ngram Corpus. The target words (in color) are shown in relation to their near associates (in gray) in the years 1800 and 2000. The meaning of Risk is shown at 11 historical points from 1800 to 2000 with a 20-year interval. PCA was performed to reduce the dimension of word embeddings from 300 to 2 so that words can be visualized in two-dimensional space. The axes represent the two principal components. A larger distance between two words indicates lower semantic similarity. The words risk, danger, and hazard started as near neighbors in 1800 but moved apart over time.

Fig. 3

Visual quantification of risk topics. (A) Heatmap of the probability that word w was generated by topic k in models derived from the Google Books Ngram Corpus (left) and the NYT Corpus (right). Words on the y-axis were selected by referring to the list of most relevant words for each topic (relevance defined by Eq. (2)) and they were grouped by categories. (B) Topic specificity (as defined by Eq. (3)). The red horizontal line indicates topic specificity equal to 1. Topics with specificity above this reference line can be considered risk-specific and therefore capture one or more aspects of the meaning of risk. Topics with specificity below 1 can be considered generic words that are not informative with respect to risk meanings. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 4

Trend analysis on risk topics derived from the Google Books Ngram Corpus. Topics are grouped into six categories: war, nuclear, health, HIV/AIDS, risk society, and economy. Relevant historical events are labeled to indicate how changes in the meanings of risk were associated with historical events and developments. Top panel: historical trends of 15 risk topics (computed using Eq. (4)). Bottom panel: normalized topic trend for each individual topic. Topic 15 is not included as it does not refer to a specific risk topic.