A writing collaboration can take many forms. One partner might be the ideas person and the other the writer. Or one might be the writer and the other the editor. Or each author might write separate story strands which are then woven together once complete. These are perfectly fine methods of collaboration but, in them, each author’s role or focus is different.
In our collaborative writing, Heide and I have been intent on being involved equally and in the same way with the project. We are both the ideas people. We are both plotters. We are both writers. We both handle all the characters and all settings. We both edit.
For example, we both placed ideas for chapter three (in which Clovenhoof forms a heavy metal band) in a form document on Dropbox. From that Heide plotted the story of Clovenhoof’s rise and fall in the music industry. I then wrote the 7,000 word chapter. Heide edited for clarity, narrative and humour. I then, based on another ideas document, plotted chapter four (in which Clovenhoof discovers he needs friends) for Heide to write. And so on we’ve gone, leapfrogging over one another.
But does it work? Or, as we asked the members of Birmingham Writers’ Group, can you tell who wrote which bits? Do we have a single authorial voice?
Feedback from our fellow writers was positive and allayed the very worst of our fears. But I wondered what computer analysis would tell us. I took a chapter written by Heide and a chapter written by myself and fed them into what text analysis websites I could find. Of course, any semi-decent experiment requires a control subject so I borrowed the text from a story by my friend, Mat Joiner, whose – SHAMELESS PLUG ALERT – work can be sampled in the collection Never Again.
Heide
|
Iain
|
Mat Joiner
|
|
Total word count :
|
3832
|
4274
|
1250
|
Number of different words :
|
1700
|
1889
|
842
|
% Words that are unique
|
44%
|
44%
|
67%
|
Readability (Gunning-Fog Index) : (6-easy 20-hard)
|
4.8
|
4.7
|
5
|
Average Syllables per Word:
|
1.54
|
1.5
|
1.47
|
Sentence count :
|
674
|
791
|
207
|
Average sentence length (words) :
|
9.67
|
9.25
|
11.23
|
Max sentence length (words) :
|
39
|
88
|
36
|
Min sentence length (words) :
|
1
|
1
|
1
|
Readability (Alternative) beta : (100-easy 20-hard, optimal 60-70)
|
67
|
70.6
|
71.5
|
So, Heide and I write sentences of comparatively similar lengths. In terms of word variety, Heide and I are identically wide-ranging in vocabulary. We are a hair’s breadth apart on the Gunning-Fog index. On the other readability scale (based on what I do not know), my words are apparently slightly more readable (more juvenile?) than Heide’s. Heide uses fractionally longer words than me.
I’m not sure if this tells us much about whether we write in a single voice. But what about word choice?
Most Common Words
|
Heide
|
Iain
|
Mat Joiner
|
1st
|
You
|
said
|
Him
|
2nd
|
clovenhoof
|
you
|
You
|
3rd
|
said
|
ben
|
Them
|
4th
|
ben
|
clovenhoof
|
Then
|
5th
|
nerys
|
nerys
|
Like
|
6th
|
Michael
|
him
|
Nettler
|
7th
|
Dave
|
man
|
What
|
8th
|
Like
|
looked
|
Face
|
9th
|
Him
|
think
|
thought
|
10th
|
Blenda
|
hand
|
Did
|
Ignoring the most common words, it turns out (shock and amazement!) we both use the words ‘said’ ‘you’ and our protagonists’ names frequently. Not much gained here.
I Write Like (www.Iwl.me) proved a little more interesting. Using algorithms and comparison methods that I am not privy to, I Write Like told me that both Heide and I write like Cory Doctorow (author of Down and Out in the Magic Kingdom) but that Mat writes like William Shakespeare. And, although I’ve no idea why any of this should be so, it is comforting to think that a piece of software thinks that Heide and I are alike in some way.
Cory Doctorow, Heide and I: Writing triplets |
The final bit of analysis I did was to see what gender each of us writes as (or should that be in). I entered our pieces of text in Gender Guesser (http://www.hackerfactor.com/GenderGuesser.php) which is a variation of the Gender Genie( http://bookblog.net/gender/genie.php). These work by looking for words that are judged to be ‘male’ or ‘female’. For example ‘with’ ‘if’ and ‘not’ are regarded as feminine whereas ‘around’ ‘what’ and ‘more’ are regarded as masculine.
And here are the results.
Heide
|
Iain
|
Mat
|
Female = 7857
|
Female = 6585
|
Female = 2063
|
Male = 8155
|
Male = 9311
|
Male = 2269
|
Difference = 298; 50.93%
|
Difference = 2726; 58.57%
|
Difference = 206; 52.37%
|
Verdict: Weak MALE
|
Verdict: Weak MALE
|
Verdict: Weak MALE
|
So, out of the three of us, I am apparently the most masculine writer, although we are all three weak male writers. The site says that ‘weak emphasis could indicate European’. I’m guessing that this means that we are judged to be European writers (which, ignoring politics, we are) although why European writers are more asexual is perhaps a question for me to ask some non-Europeans.
In conclusion, what have I learned? Probably not a lot but it’s been fun investigating.
So do we write with a single voice? I guess we won’t find out until people start reading the completed novel and tell us.