Language on usenet

This study attempts to classify the changing types of user on the internet Usenet system between 1982 and 2000, according to certain types of term used. It is not intended to be rigorously scientific.

There are five types of term, as follows.

(1) Terms most likely to be used by those educated in computer science or mathematics. These included 'recursive', 'exponential', 'boolean', 'operator', 'program', 'binary'.

(2) Terms more likely to be used by people with a general education, and less likely to be used by the uneducated. These included 'mortality', 'comprehend', 'apprehend', 'normative', 'dialectic'.

(3) Terms typically used by juveniles. These included 'retard', 'troll', 'moron', 'lame' and 'dumbest'.

(4) Terms used by cult members or conspiracy theorists, or used about them. These included 'scientology', 'fbi', 'wicca', 'objectivism'.

(5) Markers for Usenet pornography 'spammers'. These included 'cum', 'sluts', 'anal' and the word 'spam' itself.

Use of common words

Use of common words, 1982-2000. Large version here

To differentiate the general increase in use from the increase in word frequency, we took a sample of common words which anyone would be expected to use in order to communicate, irrespective of class, general knowledge or education. The words were 'without', 'every', 'most' and 'know'. Plotting the occurrence of these words by year gives chart 1, clearly suggesting that the Usenet community was stable between 1982 and 1990, and that the rapid growth did not occur until the early 1990s. There was a significant increase in 1996, probably the result of the popularisation of the internet in traditional media.

Use of key words

Relative use of key terms by type, 1982-2000. Large version here

To compute the relative occurrence of the five types of terms, we divided the number of occurrences of a given term in any year by the total number of occurrences of the common terms in the same year. For example, the word 'normative' occurs 861 times in 1992. To determine its use relative to the use of the common words, we simply divided this number by the total number of the common words, in this case 2,479,000.

To compare the use of different key terms in any given year, we took the logarithm of the increase from the 'base' ratio in 1982, to the ratio in the given year, as follows,

v = log (ratio for year/base ratio)

The values for v are shown in chart 2.

This strongly (and unsurprisingly) suggests that the early years of Usenet in the 1980s were dominated by computer scientists. It also suggests that these users fell off relative to other types of users in the 1990s.

The growth of 'generally educated' terms did not occur until the late 1980s and early 1990s, peaking around 1993, after which point use fell off. Use of 'juvenile' and 'cult' terminology grew at about the same time, but did not tail off.

The use of 'spam' terms grew fairly steadily through the whole two decades. It showed a marked increase in 1997, then an equally marked decrease in the following two years. It is difficult to say why this is, unless tools or processes were devised in that year to combat the growth in spam.

See also

• Usenet

Links

• Eternal September Note that Wikipedia, following accounts by writers such as Raymond, places 'Eternal September' in September 1993. The charts above suggest the process had begun some time before that.