Matt Rance interviewed by Aletta Stevens
What is the idea behind ProofProfessor? How did the business come about?
The idea was probably born when I set up and ran my own publishing company, Flicks Books, some years ago now, and did all the editorial tasks myself. I enjoyed that part the most, getting the text as perfect as possible. Multi-contributor books presented consistency issues – British vs. American spelling, different ways of phrasing things, e.g. ‘1950s’ or ‘fifties’, and so on – and I quickly realised I needed a system so as not to have to check the same things every time. We all know how tedious search-and-replace can be!
About four years ago I returned to working intimately with text when I began writing fiction. Back in the publishing days I had worked with WordPerfect and then Microsoft Word macros and the like, but they are size-limited and crudely DIY, and can’t be fine-tuned. I decided to create proper software to do the job. So we’ve turned the process on its head: for us, technology is the starting point for the proofreading process, rather than being an add-on at the end.
Human memory cannot hold a dictionary’s memory, which in turn cannot hold a database’s memory. The human brain is a wondrous thing but it can’t compute in the same way. And, beyond a certain age, it’s static, deteriorating. I’ll certainly acknowledge that I can’t proofread with my eyes alone. The real lever now, the game changer, as with so many things, is technology. Our software is the advantage we have over our competitors, and it enables us to keep our prices down. And in those ways it definitely seems to me an early 21st-century business!
But isn’t there a danger relying on technology?
Of course: and we don’t. But you’ll know yourself from the translation sector, with products like Trados and memoQ, that the translation process can be automised to a considerable degree. The idea with CAP (Computer-Assisted Proofreading) is similar: it’s database-driven software, with a memory. Look at Word: once upon a time it was a sensationally novel and helpful tool, but now we’ve become over-reliant on its limitations. Software is only as clever as the people who devise it, and the best, I think, is made by those who are at the coalface day after day and want a solution that will help them. That’s me.
So talk me through the nuts and bolts of the software.
OK. When a job comes in, the first thing is to run the software, which requires no human intervention at all. It trawls through the text and automatically corrects – and I must stress this – only unambiguous errors in spelling, grammar, hyphenation, punctuation, formatting, and so on, plus it alerts us (without changing the text) to things that sound odd, are inconsistent or might not be correct – maybe a word has been omitted, or a verb tense is wrong. It doesn’t make economic sense to pay a proofreader to read the document in order to find a space before a comma or correct ‘a sliver medal’ and hundreds of thousands of other language minutiae if the software can correct these things before the reading process begins. The software will also verify certain text as being correct, including whole sentences. It makes getting on for several million passes through the document.
Could you give me some specific examples?
Yes. For example, we don’t have to waste time checking the accents in ‘Raidió Teilifís Éireann’ or the spelling of ‘cephalopolysyndactyly’ or asking ourselves yet again whether ‘Stephenie Meyer’ really doesn’t contain the letter ‘a’! I want most (ideally all) of that part of the job dealt with accurately and efficiently by technology, and not by endlessly scrabbling around in dictionaries or on Google. Or worse – relying on human memory. I do see the profession as hopelessly stuck in the past: you know, all that sitting and reading with a red pen and hoping you spot everything. We’re aiming for something much more progressive, and clinical: the future of proofreading.
But do you still read the text?
Yes, of course. The manuscript is printed and proofread as normal. The changed text and the verified text are already marked, with the corrections also showing in red and Track Changes, so the client can see what we’ve done. So yes, we still do a good, old-fashioned, paper-on-desk human proofread.
You don’t proofread onscreen?
No. I find that navigating the physical environment of reading on paper is more productive: the eyes move (and tire) in a different way. I’m still a paper book man, and that dies hard. Besides, there has been some research to suggest a ‘performance deficit’ of about 25% when reading onscreen. And also that when you read on a Kindle screen, you take in less of a book than you would on paper.
As you can appreciate, the human proofreading process becomes much, much quicker: it means we can read more for flow and style and creative things the writer has done, plus spend time on any grey areas and potential pitfalls. Because our software has dealt with at least 70% of the manuscript’s crucial errors and problems, plus approved large amounts of text, we can proofread upwards of 8,000 words an hour. And finally, after completion of each job, the software is refined and enlarged, so that it’s ready for the next job.
Do you think you will want to make it licensable in the future?
Yes, I can see that happening, though not for a good while. But anyone who wants to make use of it – all they have to do is contact us. We have already started to do this with our publisher packages.
How would you respond to people who say that language is always evolving and that it doesn’t matter if we leave off an apostrophe or misspell a word?
We should all aim to write the best possible English, especially if you are an author or a publisher. That statement still needs to be said today, as, sadly, I see literacy levels plummeting – even amongst publishers and literacy bodies. It’s true that new words do appear, and one should be sensitive to changes (hyphens are disappearing, it seems to me) and take them on-board, though actually those changes are statistically minuscule and not happening as fast as people make out. Between the 2008 and 2011 COEDs, for example, a mere 400 or so words changed spelling – and that’s over three years. If the latest COED contains, say, 250,000 entries and you introduce even as many as 100 new words a year into the next edition, that represents less than 0.05% ‘evolution’! For every ‘selfie’ and ‘mumpreneur’ there are hundreds of thousands of words that don’t change from one decade or century to the next.
But if by ‘language evolving’ you mean that we should accept whatever someone’s version of language is, then I don’t agree, that’s just laziness. There has to be some kind of standard. If in doubt, reach for the COED, the gold standard, it’s that simple. Everything we do at ProofProfessor is dictionary-led. I think people too often excuse poor literacy with ‘Oh, but language is evolving’. People pay us to have their text checked to the highest level, and we’ll do that. I came across ‘prolly’ for ‘probably’ in a recent US job and that’s fine in dialogue or a casual first-person narration, but not in a PhD or an annual company report.
You’re also critical of proofreading organisations and other proofreaders. Why?
I do believe UK literacy overall is in a bad way. And it’s no longer the disadvantaged, inner-city kids stereotyped in politicians’ and TV presenters’ clichés who are failing at spelling – it’s their own teachers, county councils and politicians, their own employers. Plus journalists, professors, art gallery owners, esteemed novelists... Being in a literacy-based profession and making spelling and grammatical errors and publishing them for all to see on your website – which is now generally accepted as most people’s advertising brochure – is something I find very hard to get my head around. Why sabotage your own profile? The odd thing is that no one seems to care.
So the SfEP [Society for Editors and Proofreaders: now (12.2020) the Chartered Institute of Editing and Proofreading], of which we are not a member, has loads and loads of mistakes on its website, yet does nothing about it. What’s going on? It would be understandable if SfEP was an organisation of, say, dry-cleaners or butchers – we’d all still buy meat from a shop advertising ‘sasauges’. It seems to me a failing organisation with falling standards – and they’re all in denial.
The examinations for membership of the Institute of Translation & Interpreting are much stricter, I think.
Yes. Any Tom, Dick or Harriet can become a SfEP member. It’s very hard to burst their bubble of complacency. Aiming for the highest possible standards can alienate those who don’t. As for individual proofreaders, again I’ve offered my help (unpaid, I should add), I’ve offered to engage. But people who can’t drive very well generally don’t become driving instructors, and those who suffer from vertigo don’t become roofers. People who can’t spell properly shouldn’t really become proofreaders, should they? ‘A child of five could understand this.’ Or have I missed something?
So why not just tell them what their errors are?
Because if you see litter in your town you don’t go around picking it up: you write to the council. It’s their job (or the job of whoever drops it). Besides, ‘why join the Navy when you can be a pirate?’!
And what about the future?
In the short term, to make the software better and better, and of course to increase our client base. Simple, really... I can only see our technology becoming more relevant and necessary, certainly more exciting. We’re really only at the beginning.
See Andrew Dillon (1992). “Reading from paper versus screens: a critical review of the empirical literature”, Ergonomics, 35:10: 1297-1326.
Aletta Stevens is the author of Looking for Uncle Joop (2017) and The Remarkable Journey of Mr Prins (2020). Her articles have been published in The Author (journal of the Society of Authors), and ITI Bulletin (journal of the Institute of Translation & Interpreting). She also works as a translator: her translation of the biography of film director Paul Verhoeven is published by Faber. www.alettastevens.co.uk