Most threatning regions/countries

Data

News stories retrieved by keywords for dicplomary and military without specifing county names.

print(ndoc(toks)) # Number of news articles
## [1] 122814

Gobal and local threat

I measured the levels of threat in two ways. Selecting segmetns of texts around 10 words from threat keyword, and count frequency of country names in those. This measures perception of threat across the world (global threat; dotted line). I also pre-selected texts segments around 100 words from mentions of Japan to measure threat to Japan (local threat; solid line).

toks_jp <- tokens_select(toks, c('日本*', '東京'), window = 100)

Threat keywords

The Japanese keywords for threat perception are

hostile, danger, risk, threat, pressure, incident, use of force, emergency
print(dict[['ja']]['threat']) # Japanese words for threat
## Dictionary object with 1 key entry.
## - [threat]:
##   - 敵対*, 敵視*, 危険*, 脅威*, 威嚇*, 威圧*, 圧力, 圧迫, *有事, 武力行使, *事態

Frequency of threat keywords

This plot shows the the total number of times the threat keywords appeared in the newspaper.

mt <- dfm(toks, groups = 'year')
mt_threat <- dfm(tokens_lookup(toks, dict[['ja']]['threat']), groups = 'year', tolower = FALSE)
plot_threat(t(mt_threat))

Country names

Countries are identifies based on simple keyword matching in a hierachical geographical dictionarhy. Dictionary values for Russia contain ‘USSR’ and “Soviet *“. Japan is removed from the dictonary entirely.

Regions

Asia has been the main source of perceived threat through out the period. Global threat from Asia, which also contains the Midde East, has been higher than local threat until 2000, but there is little difference between them. While Eruope’s threat has been in declining, America became the second largest threat after 1996. Exceptions are 2008 and 2014, when Russia deployed its forces in Europe. Contraly local threat from America has been higer than global threat except the early 2000s. Europe and Africa have been percieved as source of threat to Japan than to the world for the remoteness.

mt_jp <- count_threat2(toks_jp, dict_geo, dict[['ja']]['threat'])
plot_threat(mt_jp)
mt_gl <- count_threat2(toks, dict_geo, dict[['ja']]['threat'])
plot_threat(mt_gl, TRUE)

Key sub-regions

Threat from East Asia is stronger to Japan to other counties throuh the period. The only expection is the 1996 Taiwan Strait crisis (and the following years). 2005-2015, Japan has been facing intense threat from East Asia. The threat waned in 2013-2015, but retuned in 2016. West Asia has been mainly the source of global threat, but it posed significant percieved threat to Japan in the Gulf War and the Iraq War.

dict_geo_sub <- dict_geo
names(dict_geo_sub[['ASIA']]) <-  paste(names(dict_geo_sub[['ASIA']]), 'ASIA')
names(dict_geo_sub[['EUROPE']]) <-  paste(names(dict_geo_sub[['EUROPE']]), 'EUROPE')
dict_geo_sub <- c(dict_geo_sub[['ASIA']]['EAST ASIA'], 
                  dict_geo_sub[['ASIA']]['SOUTH-EAST ASIA'],
                  dict_geo_sub[['ASIA']]['WEST ASIA'],
                  dict_geo_sub['OCEANIA'], 
                  dict_geo_sub[['EUROPE']]['EAST EUROPE'])

mt_jp <- count_threat2(toks_jp, dict_geo_sub, dict[['ja']]['threat'])
plot_threat(mt_jp)
mt_gl <- count_threat2(toks, dict_geo_sub, dict[['ja']]['threat'])
plot_threat(mt_gl, TRUE)

Key countries

At the end of the Cold War, the United States was percieved as global threat as strongly as the Soviet Union, or even strongly as local threat. The level of threat decline after the war, but continue to be the biggest threat until 2004, when China took over the position. North Korea is much less prominent than China, but the levels of perceived threat spikes time to time. After the Cold War, Russia was percieved as global threat in 2008 and 2014, but not as local threat. Despte the remoteness, Iraq was percieved as a source of threat duirng the two wars, but it is much more relevant to Japan in the 2003 Iraq War than the 1990 Gulf War.

#dict_geo_sub2 <- c(dict_geo[['ASIA']][['EAST']]['CN'],
#                  dict_geo[['ASIA']][['EAST']]['KP'],
#                  dict_geo[['ASIA']][['WEST']]['IQ'],
#                  dict_geo[['EUROPE']][['EAST']]['RU'],
#                  dict_geo[['AMERICA']][['NORTH']]['US'])

dict_geo_sub2 <- c(dict_geo[['ASIA']][['EAST']]['CN'],
                   dict_geo[['EUROPE']][['EAST']]['RU'])

mt_jp <- count_threat2(toks_jp, dict_geo_sub2, dict[['ja']]['threat'])
plot_threat(mt_jp)
mt_gl <- count_threat2(toks, dict_geo_sub2, dict[['ja']]['threat'])
plot_threat(mt_gl, TRUE)