Data Coding

Organising and classifying the input data for the generators

The input data for the generators on this site is made up of real-life material sent to real-life women in real-life contexts. It was transmitted via communications technology such as the internet, social media platforms, and mobile telephony between 1998 and the present day. New input data is added to the generators on an ongoing basis (see the Contact page of this website if you are interested in contributing material you have received yourself).

The Random Rape Threat Generator (radio edit version)

The data used in the radio edit version of the Random Rape Threat Generator is a curated selection of material from the extended remix version (see below). It has been parsed and paraphrased to help visitors quickly understand how the generator works.

The Random Rape Threat Generator (extended remix)

The complete current data sets used by the extended remix version of the generator are available on the Input Data section of the site. Below I use examples to explain how the large volume of this somewhat-amorphous text has been sub-divided into three discrete data sets that comprise the three messages displayed in the more complex version of the generator.

Wot His Lady Problem?

The Wot His Lady Problem section of the extended remix generator is made up of paraphrased material based on analysis of real-life scenarios, as well as information about targets’ contexts drawn from the various research projects described above (see ‘data set 1‘ on the Input Data page for the complete data set). Given that the motivations of individual cyberhaters are radically unknowable, there is a speculative element involved in the inclusion of entries such as ‘work is boring today’, and ‘everyone else is doing it’. Again, however, these speculations are grounded in research and analysis. The inclusion of entries such as ‘the computer is on’ and ‘it’s a Monday’, for instance, reflect my research conclusions that much of this material is sent by computer users who give little thought to: (1) what they are writing; and (2) who they are contacting and why.

Wot He Say Her?

The Wot He Say Her section of the extended remix generator is based entirely on real-life material. Archived texts have been parsed and organised into five sections (again, see the Input Data page for full details). Section one (‘content_1‘) and section two (‘content_2‘) contain various phrases of cyberhate content, for example, ‘Burn your bra and grow up’, and ‘All you do is create impotence’. Section three (‘adjective‘) comprises adjectives that antagonists use to describe women, for example, ‘bitter’, ‘crazy’, and ‘shrill’. Section four (‘noun‘) is for gendered cyberhate nouns, for example, ‘slut, ‘bitch’, and the ever-popular ‘cunt’. Section five (‘signoff‘) contains follow-up statements or sign-offs such as hashtags, emoticons, and the endlessly optimistic/creepy requests for ‘nudes’ (internet parlance for intimate photographs).

The vast bulk of the material in the Wot He Say Her section has been cut-and-pasted word-for-word from real-life contexts. However, a small amount of punctuation has been added, alongside a few instances of conjugation, adjectivisation, verbification, nominalisation, and square bracketing (where words clearly required clarification). The long and multi-staged process of transcribing this data from various locations and importing it into the extended remix version of the generator required waging all-out war with the autocorrect functions of multiple computer programs. It is possible, therefore, that, despite my best efforts, a small number of rape threats may have accidentally had a small number of spelling, grammatical, and syntax mistakes corrected. Sorry about that. For the most part, however, original presentation errors have been left intact. (It just wouldn’t be Rapeglish if a man said he would ensure his lady target would be ‘buried’ – as opposed to ‘berried’ – #10feetunder.)

While the random threats and abuse generated in the Wot He Say Her box might read bizarrely, they are an excellent representation of the weirdness that is Rapeglish. It is extremely common, for instance, for a man to issue a burlesque threat of sexual carnage before chirpily – and relatively politely – asking a woman he has never met if she will furnish photographs of her breasts and/or genitals for masturbation purposes. He may then go on to wish this woman a lovely day before signing-off with a smiley or winky face emoticon, or an ‘xxx’ (‘kiss, kiss, kiss’) or ‘xox’ (‘kiss, hug, kiss’). Determining the irony percentage of these components of Rapeglish is not always easy.

Wot He Do Next

The Wot He Do Next section (that is, ‘data set 2‘ on the Input Data page) is based only on real-life scenarios. It includes more Rapeglish, as well as some (disturbing) real-life examples of follow-up contact with and/or criminal assaults against female targets in offline contexts. The Wot He Do Next section involves paraphrased material, and has enabled the inclusion of examples of cyberhate that have been reported by targets via indirect speech. For example, in November 2011, the UK writer and activist Laurie Penny wrote a column in which she suggested that having an opinion had become ‘the short skirt of the internet’ because ‘having one and flaunting it is somehow asking an amorphous mass of almost-entirely male keyboard-bashers to tell you how they’d like to rape, kill and urinate on you’:

Efforts, too, were made to track down and harass my family, including my two school-age sisters. After one particular round of rape threats, including the suggestion that, for criticising neoliberal economic policymaking, I should be made to fellate a row of bankers at knifepoint, I was informed that people were searching for my home address. I could go on. (Penny, 2011)

Fragments from the above example appear in the Wot He Do Next data set in the following forms: ‘Tells her she should be made to fellate a row of bankers at knifepoint’; ‘Tries to track down and harass her family’; ‘Tries to track down and harass her family, including her two school-age sisters’, and ‘Tells her people are searching for her home address’. (I also included ‘A woman criticises neoliberal economic policymaking’ in the Wot His Lady Problem part of the generator.)

Parsing Rapeglish: A case study

The extended remix version of the generator includes examples of gendered cyberhate of varying length. Some particularly long messages have been inserted in full, in addition to being parsed into fragments which appear in multiple locations. To offer an illustration of this treatment of the material, consider the following message sent to the Australian writer and activist Clementine Ford:

All you do is rage rubbish you ugly duckling clementine.. When do you actually physically help anyone where you have to go out of your way with selfless service..what a dickhead as you do your pathetic best to attract attention with vulgar blabber..all you do is create impotence..now u good at that..device has integrity you know nothing about..stick to pretending to help as you are a fake and pretender..you are fat and ugly.and uneducated.. You won’t be around for long as try hard self indulgent embarrassments like you never last..bye bye joke. (cited in Ford, 2015)

Here is how this message appears in full in section one (‘content_1‘) of the Wot He Say Her data set (note that I have removed Clementine’s name and a number of extraneous full points, as well as adding some punctuation in the interests of readability):

All you do is rage rubbish, you ugly duckling. When do you actually physically help anyone where you have to go out of your way with selfless service? What a dickhead as you do your pathetic best to attract attention with vulgar blabber. All you do is create impotence. Now u good at that. Device has integrity you know nothing about. Stick to pretending to help as you are a fake and pretender. You are fat and ugly. And uneducated. You won’t be around for long as try-hard, self-indulgent embarrassments like you never last. Bye bye joke.

In addition to appearing in its entirety in the content_1 section of the Wot He Say Her data set, disaggregated fragments of this text appear elsewhere. For example, content_1 also includes the fragments ‘All you do is rage rubbish, you ugly duckling’, and ‘All you do is create impotence’ (with added capital letters, where appropriate). The adjectives ‘fat’, ‘ugly’, ‘vulgar’, ‘self-indulgent’, and ‘try-hard’, are all included in the adjective data set of Wot He Say Her (again, with the addition of capital letters and several hyphens), while the nouns ‘dickhead’, ‘embarrassment’, and ‘fake’ have been placed in the noun data set. ‘Bye bye joke’ (capitalised) also appears in the signoff data set, that is, the follow-up statement/send-off section of Wot He Say Her. Other examples of the use of fragments of this particular message can be found by referring to the Input Data page.

General generator notes

While repetition is a key feature of Rapeglish, I have deliberately held back in this regard so that users are able to access a range of different types of gendered cyberhate. If these generator were a true reflection of real life, every third comment or so would read ‘you are a fat, ugly slut’, ‘you should be raped for being so unrapeable’, and/or some variation on ‘you’re fucked but coerced sex will put you right’. As regular receivers of such material can testify, all this gets very old very fast.

Please note that an early, dead tree version of The Random Rape Threat Generator appears in Misogyny Online: A Short (and Brutish) History (Jane, 2017: 36-39).